what is maximally specific hypothesis in machine learning

Trending Categories

Selected Reading
UPSC IAS Exams Notes
Developer's Best Practices
Questions and Answers
Effective Resume Writing
HR Interview Questions
Computer Glossary

Find S Algorithm in Machine Learning

Machine learning algorithms have revolutionized the way we extract valuable insights and make informed decisions from vast amounts of data, among the multitude of algorithms, the Find-S algorithm stands out as a fundamental tool in the field. Developed by Tom Mitchell, this pioneering algorithm holds great significance in hypothesis space representation and concept learning.

With its simplicity and efficiency, the Find-S algorithm has garnered attention for its ability to discover and generalize patterns from labeled training data. In this article, we delve into the inner workings of the Find-S algorithm, exploring its capabilities and potential applications in modern machine learning paradigms.

What is the Find-S algorithm in Machine Learning?

The S algorithm, also known as the Find-S algorithm, is a machine learning algorithm that seeks to find a maximally specific hypothesis based on labeled training data. It starts with the most specific hypothesis and generalizes it by incorporating positive examples. It ignores negative examples during the learning process.

The algorithm's objective is to discover a hypothesis that accurately represents the target concept by progressively expanding the hypothesis space until it covers all positive instances.

Symbols used in Find-S algorithm

In the Find-S algorithm, the following symbols are commonly used to represent different concepts and operations −

∅ (Empty Set) − This symbol represents the absence of any specific value or attribute. It is often used to initialize the hypothesis as the most specific concept.

? (Don't Care) − The question mark symbol represents a "don't care" or "unknown" value for an attribute. It is used when the hypothesis needs to generalize over different attribute values that are present in positive examples.

Positive Examples (+) − The plus symbol represents positive examples, which are instances labeled as the target class or concept being learned.

Negative Examples (-) − The minus symbol represents negative examples, which are instances labeled as non-target classes or concepts that should not be covered by the hypothesis.

Hypothesis (h) − The variable h represents the hypothesis, which is the learned concept or generalization based on the training data. It is refined iteratively throughout the algorithm.

These symbols help in representing and manipulating the hypothesis space and differentiating between positive and negative examples during the hypothesis refinement process. They aid in capturing the target concept and generalizing it to unseen instances accurately.

Inner working of Find-S algorithm

The Find-S algorithm operates on a hypothesis space to find a general hypothesis that accurately represents the target concept based on labeled training data. Let's delve into the inner workings of the algorithm −

Initialization − The algorithm starts with the most specific hypothesis, denoted as h. This initial hypothesis is the most restrictive concept and typically assumes no positive examples. It may be represented as h = <∅, ∅, ..., ∅>, where ∅ denotes "don't care" or "unknown" values for each attribute.

Iterative Process − The algorithm iterates through each training example and refines the hypothesis based on whether the example is positive or negative.

For each positive training example (an example labeled as the target class), the algorithm updates the hypothesis by generalizing it to include the attributes of the example. The hypothesis becomes more general as it covers more positive examples.

For each negative training example (an example labeled as a non-target class), the algorithm ignores it as the hypothesis should not cover negative examples. The hypothesis remains unchanged for negative examples.

Generalization − After processing all the training examples, the algorithm produces a final hypothesis that covers all positive examples while excluding negative examples. This final hypothesis represents the generalized concept that the algorithm has learned from the training data.

During the iterative process, the algorithm may introduce "don't care" symbols or placeholders (often denoted as "?") in the hypothesis for attributes that vary among positive examples. This allows the algorithm to generalize the concept by accommodating varying attribute values. The algorithm discovers patterns in the training data and provides a reliable representation of the concept being learned.

Let's explore the steps of the algorithm using a practical example −

Suppose, we have a dataset of animals with two attributes: "has fur" and "makes sound." Each animal is labeled as either a dog or a cat. Here is a sample training dataset −

Animal	Has Fur	Makes Sound	Label
Dog	Yes	Yes	Dog
Cat	Yes	No	Cat
Dog	No	Yes	Dog
Cat	No	No	Cat
Dog	Yes	Yes	Dog

To apply the Find-S algorithm, we start with the most specific hypothesis, denoted as h, which initially represents the most restrictive concept. In our example, the initial hypothesis would be h = <∅, ∅>, indicating that no specific animal matches the concept.

For each positive training example (an example labeled as the target class), we update the hypothesis h to include the attributes of that example. In our case, the positive training examples are dogs. Therefore, h would be updated to h = <Yes, Yes>.

For each negative training example (an example labeled as a non-target class), we ignore it as the hypothesis h should not cover those examples. In our case, the negative training examples are cats, and since h already covers dogs, we don't need to update the hypothesis.

After processing all the training examples, we obtain a generalized hypothesis that covers all positive training examples and excludes negative examples. In our example, the final hypothesis h = <Yes, Yes> accurately represents the concept of a dog.

Here is a Python program illustrating the Find-S algorithm −

In this program, the training data is represented as a list of tuples. The algorithm iterates through each example, updating the hypothesis accordingly. The final hypothesis represents the concept of a dog based on the training data.

The Find-S algorithm serves as a foundation for more complex machine learning algorithms and has practical applications in various domains, including classification, pattern recognition, and decision-making systems.

In conclusion, the Find-S algorithm has proven to be a powerful tool in machine learning, allowing us to learn concepts and generalize patterns from labeled training data. With its iterative process and ability to find maximally specific hypotheses, this algorithm has paved the way for advancements in hypothesis space representation and concept learning, making it a fundamental technique in the field. Its simplicity and effectiveness make it a valuable asset in various machine learning applications.

Related Articles
Understanding node2vec algorithm in machine learning
Are machine learning algorithm patentable?
Statistical Comparison of Machine Learning Algorithm
Mini Batch K-means clustering algorithm in Machine Learning
How to Choose the right Machine Learning algorithm?
What is a Simple Genetic Algorithm (SGA) in Machine Learning?
What are different components of a machine learning algorithm?
How to Explain Steady State Genetic Algorithm (SSGA) in Machine Learning?
Everything to know about Amazon\'s new Machine Learning Platform
Machine Learning – The Intelligent Machine
TabNet in Machine Learning
Clustering in Machine Learning
Using learning curves in Machine Learning Explained
Roadmap to study AI, Machine Learning, and Deep Machine Learning
Demystifying Machine Learning

Kickstart Your Career

Get certified by completing the course

cryptocurrency
Digital Marketing

Discover the Power of Find S Algorithm – A Comprehensive Guide

Do you struggle to find the right algorithm for your machine-learning tasks?

It can be frustrating and time-consuming to search through countless options and experiment with different algorithms, hoping to find the one that fits your data and requirements.

Fortunately, there is a powerful and widely used algorithm in machine learning called the “find s algorithm” that can help you automate the process of finding a suitable hypothesis for your data.

In this article, we will explore the find s algorithm, how it works, and how you can use it to improve your machine learning workflows.

Whether you are a beginner or an experienced practitioner, this guide will provide valuable insights and practical tips for leveraging the find-s algorithm to achieve better results and save time.

So, let’s dive in!

What is find s algorithm in machine learning?

Discover the Power of Find S Algorithm - A Comprehensive Guide

The Find-S algorithm is a fundamental technique in machine learning that aims to discover a generalized hypothesis from a given set of training data.

It is commonly used in concept learning tasks, where the goal is to learn a concept or rule from a set of positive and negative examples.

The Find-S algorithm follows a simple and intuitive approach. It starts with the most specific hypothesis, represented by a boundary set of attribute-value pairs that can classify the positive examples correctly.

As it iterates through the training data, the algorithm generalizes the hypothesis by specializing the boundary set whenever it encounters a negative example.

During each iteration, the Find-S algorithm compares the attributes of the current positive example with the boundary set.

If an attribute-value pair in the boundary set contradicts the positive example, the algorithm narrows down the boundary by replacing the contradictory attribute-value pair with a more specific one.

The algorithm continues this process until it traverses all the positive examples and builds a generalized hypothesis that can accurately classify unseen examples.

The resulting hypothesis represents the most general concept that satisfies the training data.

The Find-S algorithm’s simplicity and efficiency make it an effective technique for concept learning in machine learning.

It provides a foundation for more advanced algorithms and serves as a stepping stone in understanding the intricacies of hypothesis generation and concept generalization.

Find s algorithm in machine learning with an example

The Find-S algorithm is a popular technique in machine learning that aids in concept learning and hypothesis generation. It efficiently discovers a generalized hypothesis from a set of positive training examples. Let’s illustrate the workings of the Find-S algorithm through a simple example.

Consider a task where we aim to learn a concept of “fruit” based on attributes like shape, color, and taste. We have a training dataset with positive examples of apples, oranges, and bananas.

The Find-S algorithm starts with an initial hypothesis that represents the most specific concept. For instance, it may begin with a hypothesis like “fruit has shape ? and color ? and taste ?.”

As the algorithm iterates through the positive examples, it updates the hypothesis by generalizing it. Suppose the first positive example is an apple with attributes (shape: round, color: red, taste: sweet). The algorithm modifies the hypothesis to “fruit has shape round and color red and taste sweet.”

For the next positive example, let’s say it’s an orange with attributes (shape: round, color: orange, taste: sour). The algorithm further generalizes the hypothesis to “fruit has shape round and color ? and taste ?.”

Finally, when encountering a positive example of a banana with attributes (shape: elongated, color: yellow, taste: sweet), the algorithm generalizes the hypothesis to “fruit has shape ? and color ? and taste sweet.”

At the end of the algorithm, we obtain a generalized hypothesis that accurately describes the concept of “fruit” based on the provided positive examples.

The Find-S algorithm is a valuable tool in machine learning, allowing us to learn concepts from limited training data and generalize them effectively. Its simplicity and effectiveness make it a cornerstone technique in concept learning tasks.

Here’s an example to illustrate how FIND S algorithm works:-

Suppose we want to build a machine learning model to identify a type of flower based on its petal color and size.

We have a set of training data containing examples of flowers along with their attributes, as shown below:-

Petal Color	Petal Size	Flower Type
Red	Small	Rose
Blue	Small	Bluebell
Red	Large	Lily
Blue	Small	Bluebell
Red	Small	Rose
Blue	Large	Bluebell

The FIND-S algorithm works by initializing the most specific hypothesis, S , to be the set of all possible hypotheses. In this case, S is the set of all possible combinations of petal color and size:-

S = {<Red, Small>, <Red, Large>, <Blue, Small>, <Blue, Large>}

The algorithm then iterates over the training examples, and for each positive example, it updates S to include only the attributes that match the example.

For instance, the first positive example is a Rose with red and small petals. Therefore, we can update S to include only the attributes that match this example:-

S = {<Red, Small>}

Next, the algorithm considers the second positive example, which is a Bluebell with small petals. We can again update S to include only the attributes that match this example:-

S = {<Red, Small>, <Blue, Small>}

The algorithm continues to iterate over the remaining training examples, updating S for each positive example. Eventually, the final hypothesis for S becomes:-

This hypothesis is the most specific hypothesis that fits all positive examples in the training data. It can predict the flower type of new examples based on their petal color and size.

Find s Algorithm advantages and disadvantages.

The Find-S algorithm is a valuable tool in machine learning for concept learning tasks. Like any algorithm, it comes with its own set of advantages and disadvantages.

Let’s explore them in detail.

Advantages: –

Simplicity : The Find-S algorithm is simple and easy to understand, making it accessible to beginners in machine learning.

Efficiency : It can efficiently generate a generalized hypothesis by iterating through the positive training examples, reducing the computational complexity.

Interpretability : The generated hypothesis is human-readable and interpretable, providing insights into the learned concept.

Incremental learning : The algorithm accommodates incremental learning, allowing the addition of new training examples without retraining the entire model.

Disadvantages: –

Limited expressiveness : The Find-S algorithm assumes a restricted hypothesis space, leading to limited representation power for complex concepts.

Sensitivity to noise : It is sensitive to noisy or incorrect data. In the presence of outliers or mislabeled examples, the algorithm may generate an inaccurate hypothesis.

Lack of negative examples : The algorithm heavily relies on positive examples, lacking explicit negative examples for learning.

Restriction to attribute-value representation : The algorithm assumes a fixed attribute-value representation, making it less suitable for handling continuous or complex data.

Here’s a table summarizing the advantages and disadvantages of the Find-S algorithm:-

Advantages	Disadvantages
Simple and easy to implement	Can only represent simple hypotheses
Guaranteed to converge to a consistent hypothesis	May not find the most accurate hypothesis
Can handle noisy data	It does not take into account prior knowledge or background information
Requires very little memory and computation	Cannot handle continuous or non-categorical data
Works well with small to medium-sized datasets	Sensitive to the order of training examples
It can be used with different types of classification problems	Requires labeled training data
You can get stuck in local optima	Does not take into account prior knowledge or background information

Limitations of find s algorithm

While the Find-S algorithm is a useful technique in machine learning , it does have certain limitations that need to be considered when applying it to real-world scenarios.

Understanding these limitations can help researchers and practitioners make informed decisions about its usage.

One major limitation of the Find-S algorithm is its restrictive hypothesis space . The algorithm assumes a specific representation, such as attribute-value pairs, which may not be suitable for complex or continuous data.

This limitation can hinder its effectiveness in handling diverse and nuanced concepts.

Another limitation is its sensitivity to noise and outliers . The algorithm heavily relies on the training data, and any inaccuracies or mislabeled examples can lead to an incorrect or overly specific hypothesis.

This sensitivity to noise can impact the generalizability and robustness of the learned concept.

The Find-S algorithm also has limited expressiveness . It struggles with capturing complex relationships or patterns that may exist within the data.

This limitation makes it less effective in scenarios where more sophisticated models or algorithms are required to learn intricate concepts.

Furthermore, the algorithm assumes that negative examples are absent . This can be problematic as negative examples are crucial for differentiating between concept boundaries, leading to potential errors or biases in the learned hypothesis.

Despite these limitations, the Find-S algorithm serves as a foundational tool in concept learning.

It offers simplicity and efficiency, making it suitable for basic applications, but it may require enhancements or alternative algorithms to address its limitations when dealing with more complex datasets or concepts.

here’s a table outlining some of the limitations of the Find-S algorithm:-

Limitations of Find-S Algorithm	Explanation
Limited to Binary Hypothesis Space	The Find-S algorithm can only work with binary (true/false) hypothesis spaces, meaning that it cannot handle more complex or continuous data sets.
Assumes Consistency of Data	The algorithm assumes that the data provided is consistent, meaning that there are no conflicting examples that cannot be classified into a single hypothesis. If the data is inconsistent, the algorithm may not produce a correct hypothesis.
Cannot Handle Noise	The Find-S algorithm cannot handle noisy data, meaning that if the data contains errors or outliers, the resulting hypothesis may be incorrect.
Limited to Concept Learning	The Find-S algorithm is limited to concept learning, which means that it can only learn to classify data based on a set of predefined categories. It cannot learn to recognize more complex patterns or relationships in the data.
May Produce Overly Specific Hypotheses	The Find-S algorithm may produce hypotheses that are too specific and only work for the training data set, but fail to generalize to new, unseen data. This is known as overfitting.
May Require a Large Training Set	The Find-S algorithm may require a large number of training examples to produce an accurate hypothesis, especially if the data is complex or noisy.

Difference between find-s and candidate elimination algorithm

When it comes to concept learning in machine learning , two notable algorithms that are often employed are the Find-S algorithm and the Candidate Elimination algorithm .

While they share similarities in their objective of generating hypotheses, they differ in their approaches and functionality.

The main difference between the Find-S algorithm and the Candidate Elimination algorithm lies in their hypothesis representation.

The Find-S algorithm generates the most specific hypothesis that covers all positive training examples. It starts with the most specific hypothesis and generalizes it iteratively as it encounters positive examples.

On the other hand, the Candidate Elimination algorithm generates the most general and most specific hypotheses simultaneously.

It maintains a set of hypotheses and updates them based on positive and negative training examples. The algorithm eliminates hypotheses that are inconsistent with the observed data while retaining the general and specific boundaries.

Another distinction lies in their handling of negative examples . The Find-S algorithm does not explicitly consider negative examples during the learning process, focusing solely on positive instances.

In contrast, the Candidate Elimination algorithm incorporates negative examples to refine the hypothesis space and narrow down the possible solutions.

Furthermore, the Candidate Elimination algorithm allows for incremental learning , accommodating new training examples as they arrive without the need for retraining the entire model. This adaptability makes it suitable for dynamic environments and evolving datasets.

In summary, the Find-S algorithm generates the most specific hypothesis based on positive examples, while the Candidate Elimination algorithm maintains and refines both the most general and most specific hypotheses, incorporating both positive and negative examples.

Their differing approaches make each algorithm suitable for different learning scenarios and requirements.

Here is a table highlighting the key differences between the Find-S algorithm and the Candidate Elimination algorithm:-

Algorithm	Find-S	Candidate Elimination
Input	A set of positive and negative training examples.	A set of training examples and a language space.
Output	The most specific hypothesis that fits all positive training examples.	The set of hypotheses that fit all positive training examples and exclude all negative training examples.
Hypothesis Space	Consists of only the most specific hypothesis (initialized as the most specific hypothesis).	Consists of all hypotheses that can be expressed in the language space.
Search Strategy	Start with the most specific hypothesis and generalize it as necessary to fit positive training examples.	Start with the most general hypothesis and specialize it as necessary to exclude negative training examples.
Completeness	Guaranteed to find a solution if one exists in the hypothesis space.	Guaranteed to find a solution if one exists in the language space.
Efficiency	Can converge quickly if the hypothesis space is small.	May take longer to converge if the language space is large.
Robustness	May overfit the training data if the hypothesis space is too small.	Can handle noise in the training data by considering multiple hypotheses.
Limitations	Limited to finding the most specific hypothesis in the hypothesis space.	Can find a set of solutions that fit all training examples, but may not find a unique solution.

Why do we use the Find-S algorithm?

The Find-S algorithm is a popular tool in machine learning that finds utility in various scenarios. Its simplicity and effectiveness make it a valuable choice for certain concept learning tasks.

Let’s delve into the reasons why the Find-S algorithm is commonly used.

1. Simplicity : The Find-S algorithm is straightforward to implement and comprehend. Its simplicity makes it accessible to beginners and serves as a foundation for understanding more complex machine learning techniques.

2. Efficiency : The algorithm operates efficiently, especially when dealing with small or well-defined concept spaces. It can generate a generalized hypothesis by iterating through positive training examples, reducing computational complexity.

3. Interpretability : The hypotheses generated by the Find-S algorithm are human-readable and interpretable. This attribute provides insights into the learned concept and facilitates domain experts’ understanding and decision-making.

4. Incremental learning : The algorithm can accommodate incremental learning, allowing the addition of new training examples without retraining the entire model. This flexibility makes it suitable for dynamic environments where data evolves over time.

5. Initial hypothesis : The Find-S algorithm starts with the most specific hypothesis, providing a solid starting point for further refinement or exploration. This property enables the algorithm to converge quickly to a reasonable hypothesis.

By leveraging the advantages of simplicity, efficiency, interpretability, incremental learning, and an appropriate initial hypothesis, the Find-S algorithm remains a valuable choice for concept learning tasks.

While it may have limitations, its practical benefits make it a go-to approach in certain machine-learning scenarios.

here is a table on why we use the Find-S algorithm:-

Reason	Explanation
Automating Hypothesis Formation	The Find-S algorithm is used to automate the process of hypothesis formation in machine learning. Given a set of training data, it can generate a hypothesis that can predict the class labels of unseen examples.
Concept Learning	The Find-S algorithm is used in concept learning to find the most specific hypothesis that is consistent with the training data. This hypothesis can be used to classify new examples as belonging to a particular concept or not.
Simplifying the Hypothesis Space	The Find-S algorithm simplifies the hypothesis space by only considering hypotheses that are consistent with the training data. This reduces the search space and makes the learning process more efficient.
Handling Noisy Data	The Find-S algorithm can handle noisy data by finding the most specific hypothesis that is consistent with the noisy data. This helps to reduce the impact of noise on the learning process.
Interpretability	The Find-S algorithm generates hypotheses that are easily interpretable by humans. This makes it useful in applications where the interpretability of the learned model is important, such as in medical diagnosis or legal decision-making.

What is the output obtained by Find-S algorithm?

The Find-S algorithm is a concept learning technique in machine learning that aims to generate a generalized hypothesis from a set of positive training examples.

The output obtained by the Find-S algorithm is a specific hypothesis that accurately represents the learned concept based on the provided positive instances.

The specific hypothesis generated by the Find-S algorithm is typically in the form of an attribute-value representation .

It describes the boundaries and constraints of the learned concept by specifying the values of different attributes that define the concept.

For example, suppose we are using the Find-S algorithm to learn the concept of a “bird” based on positive training examples of different bird species. The output hypothesis might be something like “Bird has wings: true, beak: true, feathers: true, and can fly: true.”

The output obtained by the Find-S algorithm is tailored to the specific positive training examples, ensuring that it covers all the provided instances while remaining as specific as possible.

It represents the most specific generalization of the concept based on the available information.

The specific hypothesis produced by the Find-S algorithm can be used for further classification of new, unseen examples.

It serves as a learned model that can categorize instances into the concept it represents, aiding in decision-making and prediction tasks.

In summary, the output obtained by the Find-S algorithm is a specific hypothesis that defines the learned concept based on the positive training examples provided.

It represents the boundaries and attributes that characterize the concept of interest.

What is the algorithm for finding a maximally specific hypothesis?

The process of finding a maximally specific hypothesis is a fundamental step in machine learning for concept learning tasks.

This algorithm allows us to generate the most specific hypothesis that fits the given positive training examples. Let’s delve into the steps involved in this process.

The algorithm starts with an empty hypothesis , typically denoted as h0 , which represents the most general hypothesis.

It contains a set of n attributes, each initialized to a special value like “null” or “?”.

For each positive training example, the algorithm examines the attributes’ values.

If an attribute in h0 is already assigned a specific value and it contradicts the current positive example, the algorithm updates the value to a more general one, such as “?”. If the attribute is unassigned or consistent with the positive example, it remains unchanged.

The algorithm iterates through all the positive training examples, updating the attribute values in h0 as necessary.

After considering all the examples, the resulting hypothesis, denoted as h , represents the maximally specific hypothesis. It is the most specific hypothesis that can classify all the positive training examples accurately.

The algorithm works by refining the hypothesis based on the available positive instances, gradually narrowing down the attribute values to fit the positive examples.

It ensures that the maximally specific hypothesis is obtained, capturing the specific boundaries and constraints of the concept based on the provided training data.

In summary, the algorithm for finding a maximally specific hypothesis initializes an empty hypothesis and iteratively refines it based on the positive training examples.

It generates the most specific hypothesis that accurately represents the concept and covers all the positive instances.

The steps of the Specific-to-General Algorithm are as follows:-

Initialize the hypothesis space to include only the most specific hypothesis possible.
For each positive training example, remove from the hypothesis space any hypothesis that does not include that example.
For each negative training example, remove from the hypothesis space any hypothesis that includes that example.
Generalize the remaining hypotheses by removing any attribute values that are not present in any positive training example.
Continue Steps 2-4 until there are no more changes to the hypothesis space.
Return the maximally specific hypothesis, if it exists.

The Specific-to-General Algorithm is a simple and efficient way to learn a maximally specific hypothesis from a set of training examples.

It is widely used in machine learning and has applications in many different domains, such as natural language processing, computer vision, and robotics.

How does find s algorithm start from the most specific hypothesis and generalize it?

It starts with the most specific hypothesis possible and generalizes it as it encounters more examples.

The algorithm starts by initializing the hypothesis to the most specific hypothesis possible, which in the case of a binary classification problem is a hypothesis that classifies all instances as negative.

This hypothesis is represented as a conjunction of literals that describe the attributes of an instance. For example, in a problem where we want to classify whether a person is a student or not based on their age and enrollment status, the most specific hypothesis would be:

“Not a student AND Age = 0”

As the algorithm encounters positive examples, it updates the hypothesis to include the attributes that are present in all positive examples. For example, if it encounters an example of a student who is 20 years old and enrolled, the hypothesis would be updated to:

“Student AND Age = 20 AND Enrollment = True”

If the algorithm encounters a negative example, it does not update the hypothesis. Instead, it moves on to the next example.

The algorithm continues to update the hypothesis with each positive example it encounters, generalizing it to include more attributes that are common to all positive examples.

Eventually, the hypothesis will generalize to include all attributes that are common to all positive examples and no attributes that are common to negative examples.

In the end, the algorithm will output the final hypothesis, which represents the set of attributes that best describe the positive examples and discriminates them from the negative examples. This final hypothesis can then be used to classify new instances.

In conclusion, understanding the Find-S algorithm can be crucial for machine learning enthusiasts and professionals alike.

With its ability to generalize a set of hypotheses and reduce complexity, the algorithm is widely used in various fields, including data mining, artificial intelligence, and computer science.

As we have seen, the algorithm operates by iteratively eliminating hypotheses that are not consistent with the data, until it finds the most specific hypothesis that fits all the positive examples.

By using this powerful tool, you can enhance your predictive modeling skills and generate accurate insights from your data.

So, if you want to master the Find-S algorithm, start exploring its applications and experimenting with its implementation.

Project Management Skills You Need to Succeed in an IT Company

what is maximally specific hypothesis in machine learning

Your Guide to Starting a Small Business as a Solopreneur

Learn the Secrets of Winning on Twitter

Affiliate Disclosure

Data Science Introduction

What Is Data Science? A Beginner's Guide To Data Science
Data Science Tutorial – Learn Data Science from Scratch!
What are the Best Books for Data Science?
Top 15 Hot Artificial Intelligence Technologies
Top 8 Data Science Tools Everyone Should Know
Top 10 Data Analytics Tools You Need To Know In 2024
5 Data Science Projects – Data Science Projects For Practice
Top 10 Data Science Applications with Real Life Examples in 2024
Who is a Data Scientist?
SQL For Data Science: One stop Solution for Beginners

Statistical Inference

All You Need To Know About Statistics And Probability
A Complete Guide To Math And Statistics For Data Science
Introduction To Markov Chains With Examples – Markov Chains With Python
What is Fuzzy Logic in AI and What are its Applications?
How To Implement Bayesian Networks In Python? – Bayesian Networks Explained With Examples
All You Need To Know About Principal Component Analysis (PCA)
Python for Data Science – How to Implement Python Libraries

Machine Learning

What is Machine Learning? Machine Learning For Beginners
Which is the Best Book for Machine Learning?
Mathematics for Machine Learning: All You Need to Know
Top 10 Machine Learning Frameworks You Need to Know

Predicting the Outbreak of COVID-19 Pandemic using Machine Learning

Introduction to machine learning: all you need to know about machine learning.

Machine Learning Tutorial for Beginners

Top 10 Applications of Machine Learning in Daily Life

Machine learning algorithms, how to implement find-s algorithm in machine learning.

What is Cross-Validation in Machine Learning and how to implement it?
All You Need To Know About The Breadth First Search Algorithm

Supervised Learning

What is Supervised Learning and its different types?
Linear Regression Algorithm from Scratch
How To Implement Linear Regression for Machine Learning?
Introduction to Classification Algorithms
How To Implement Classification In Machine Learning?
Naive Bayes Classifier: Learning Naive Bayes with Python
A Comprehensive Guide To Naive Bayes In R
A Complete Guide On Decision Tree Algorithm
Decision Tree: How To Create A Perfect Decision Tree?
What is Overfitting In Machine Learning And How To Avoid It?
How To Use Regularization in Machine Learning?

Unsupervised Learning

What is Unsupervised Learning and How does it Work?
K-means Clustering Algorithm: Know How It Works
KNN Algorithm: A Practical Implementation Of KNN Algorithm In R
Implementing K-means Clustering on the Crime Dataset
K-Nearest Neighbors Algorithm Using Python
Apriori Algorithm : Know How to Find Frequent Itemsets
What Are GANs? How and why you should use them!
Q Learning: All you need to know about Reinforcement Learning

Miscellaneous

Data Science vs Machine Learning - What's The Difference?
AI vs Machine Learning vs Deep Learning
Data Analyst vs Data Engineer vs Data Scientist: Salary, Skills, Responsibilities
Data Science vs Big Data vs Data Analytics

Career Opportunities

Data Science Career Opportunities: Your Guide To Unlocking Top Data Scientist Jobs
Data Scientist Skills – What Does It Take To Become A Data Scientist?
10 Skills To Master For Becoming A Data Scientist
Data Scientist Resume Sample – How To Build An Impressive Data Scientist Resume
Data Scientist Salary – How Much Does A Data Scientist Earn?
Machine Learning Engineer vs Data Scientist : Career Comparision
How To Become A Machine Learning Engineer? – Learning Path

Interview Questions

Top Machine Learning Interview Questions You Must Prepare In 2024
Top Data Science Interview Questions For Budding Data Scientists In 2024
120+ Data Science Interview Questions And Answers for 2024

Artificial Intelligence

In Machine Learning , concept learning can be termed as “ a problem of searching through a predefined space of potential hypothesis for the hypothesis that best fits the training examples” – Tom Mitchell. In this article, we will go through one such concept learning algorithm known as the Find-S algorithm. If you want to go beyond this article and really want the level of expertise in you – you can get certified in Machine Learning Certification!

Machine Learning Full Course – Learn Machine Learning 10 Hours | Machine Learning Tutorial | Edureka

Machine Learning Course lets you master the application of AI with the expert guidance. It includes various algorithms with applications.

The following topics are discussed in this article.

What is Find-S Algorithm in Machine Learning?

How Does it Work?

Limitations of Find-S Algorithm

Implementation of find-s algorithm.

In order to understand Find-S algorithm, you need to have a basic idea of the following concepts as well:

Concept Learning
General Hypothesis
Specific Hypothesis

1. Concept Learning

Let’s try to understand concept learning with a real-life example. Most of human learning is based on past instances or experiences. For example, we are able to identify any type of vehicle based on a certain set of features like make, model, etc., that are defined over a large set of features.

These special features differentiate the set of cars, trucks, etc from the larger set of vehicles. These features that define the set of cars, trucks, etc are known as concepts.

Similar to this, machines can also learn from concepts to identify whether an object belongs to a specific category or not. Any algorithm that supports concept learning requires the following:

Training Data
Target Concept
Actual Data Objects

2. General Hypothesis

Hypothesis, in general, is an explanation for something. The general hypothesis basically states the general relationship between the major variables. For example, a general hypothesis for ordering food would be I want a burger.

G = { ‘?’, ‘?’, ‘?’, …..’?’}

3. Specific Hypothesis

The specific hypothesis fills in all the important details about the variables given in the general hypothesis. The more specific details into the example given above would be I want a cheeseburger with a chicken pepperoni filling with a lot of lettuce.

S = {‘Φ’,’Φ’,’Φ’, ……,’Φ’}

Now ,let’s talk about the Find-S Algorithm in Machine Learning.

The Find-S algorithm follows the steps written below:

Initialize ‘h’ to the most specific hypothesis.
The Find-S algorithm only considers the positive examples and eliminates negative examples. For each positive example, the algorithm checks for each attribute in the example. If the attribute value is the same as the hypothesis value, the algorithm moves on without any changes. But if the attribute value is different than the hypothesis value, the algorithm changes it to ‘?’.

Now that we are done with the basic explanation of the Find-S algorithm, let us take a look at how it works.

How Does It Work?

The process starts with initializing ‘h’ with the most specific hypothesis, generally, it is the first positive example in the data set.
We check for each positive example. If the example is negative, we will move on to the next example but if it is a positive example we will consider it for the next step.
We will check if each attribute in the example is equal to the hypothesis value.
If the value matches, then no changes are made.
If the value does not match, the value is changed to ‘?’.
We do this until we reach the last positive example in the data set.

There are a few limitations of the Find-S algorithm listed down below:

There is no way to determine if the hypothesis is consistent throughout the data.
Inconsistent training sets can actually mislead the Find-S algorithm, since it ignores the negative examples.
Find-S algorithm does not provide a backtracking technique to determine the best possible changes that could be done to improve the resulting hypothesis.

Top 10 Trending Technologies to Learn in 2024 | Edureka

Now that we are aware of the limitations of the Find-S algorithm, let us take a look at a practical implementation of the Find-S Algorithm.

To understand the implementation, let us try to implement it to a smaller data set with a bunch of examples to decide if a person wants to go for a walk.

The concept of this particular problem will be on what days does a person likes to go on walk.


Morning	Sunny	Warm	Yes	Mild	Strong	Yes
Evening	Rainy	Cold	No	Mild	Normal	No
Morning	Sunny	Moderate	Yes	Normal	Normal	Yes
Evening	Sunny	Cold	Yes	High	Strong	Yes

Looking at the data set, we have six attributes and a final attribute that defines the positive or negative example. In this case, yes is a positive example, which means the person will go for a walk.

So now, the general hypothesis is:

h 0 = {‘Morning’, ‘Sunny’, ‘Warm’, ‘Yes’, ‘Mild’, ‘Strong’}

This is our general hypothesis, and now we will consider each example one by one, but only the positive examples.

h 1 = {‘Morning’, ‘Sunny’, ‘?’, ‘Yes’, ‘?’, ‘?’}

h 2 = {‘?’, ‘Sunny’, ‘?’, ‘Yes’, ‘?’, ‘?’}

We replaced all the different values in the general hypothesis to get a resultant hypothesis. Now that we know how the Find-S algorithm works, let us take a look at an implementation using Python .

Let’s try to implement the above example using Python . The code to implement the Find-S algorithm using the above data is given below.

This brings us to the end of this article where we have learned the Find-S Algorithm in Mach ine Learning with its implementation and use case. I hope you are clear with all that has been shared with you in this tutorial.

With immense applications and easier implementations of Python with data science, there has been a significant increase in the number of jobs created for data science every year. Enroll for Edureka’s Data Science with Python and get hands-on experience with real-time industry projects along with 24×7 support, which will set you on the path of becoming a successful Data Scientist,

We are here to help you with every step on your journey and come up with a curriculum that is designed for students and professionals who want to be a Machine Learning Engineer . The course is designed to give you a head start into Python programming and train you for both core and advanced Python concepts along with various Machine learning Algorithms like SVM , Decision Tree , etc.

If you come across any questions, feel free to ask all your questions in the comments section of “Find-S Algorithm In Machine Learning” and our team will be glad to answer.

","eventStatus":"https://schema.org/EventScheduled","image":"https://d1jnx9ba8s6j9r.cloudfront.net/imgver.0000000000/img/co_img_2263_1678870066.png","startDate":"2024-09-07T07:00:00+0000","endDate":"2024-09-22T10:00:00+0000","Duration":"R8/P7DT4H","inLanguage":"en-US","url":"https://www.edureka.co/masters-program/machine-learning-engineer-training?utm_source=blogbatch&utm_campaign=batch_details","eventAttendanceMode":"https://schema.org/OnlineEventAttendanceMode","location":[{"@type":"VirtualLocation","url":"https://www.edureka.co/masters-program/machine-learning-engineer-training?utm_source=blogbatch&utm_campaign=batch_details"}],"organizer":{"@type":"Thing","name":"Edureka","url":"https://www.edureka.co"},"performer":"Edureka","offers":{"@type":"AggregateOffer","url":"https://www.edureka.co/masters-program/machine-learning-engineer-training?utm_source=blogbatch&utm_campaign=batch_details","availability":"InStock","price":"1499","priceCurrency":"USD","lowprice":"1499","highprice":"1499","validFrom":"2024-09-04T00:00:00+0000"}}

Course Name	Date	Details
	Class Starts on 7th September,2024 7th September SAT&SUN (Weekend Batch)

Course Name

Date

Details

Class Starts on 7th September,2024

7th September

SAT&SUN (Weekend Batch)

ChatGPT Training Course: Beginners to Advance ...

15k Enrolled Learners

Artificial Intelligence Certification Course

16k Enrolled Learners

Prompt Engineering Course

5k Enrolled Learners

Generative AI in Business: University of Camb ...

1k Enrolled Learners

MLOps Certification Course Online

6k Enrolled Learners

Large Language Models (LLMs) Certification Co ...

3k Enrolled Learners

Reinforcement Learning

Graphical models certification training, generative ai in hr certification course, browse categories, subscribe to our newsletter, and get personalized recommendations..

Already have an account? Sign in .

20,00,000 learners love us! Get personalised resources in your inbox.

At least 1 upper-case and 1 lower-case letter

Minimum 8 characters and Maximum 50 characters

We have recieved your contact details.

You will recieve an email from us shortly.

+(1) 647-467-4396
hello@knoldus.com

hello@knoldus.com

Concept Learning: The stepping stone towards Machine Learning with Find-S

From our previous blog , we came across what awesome stuff a machine can do with machine learning and what all math stuff is required before you take a deep dive into machine learning. Now we all know the prerequisites for machine learning, so let’s start the journey towards machine learning with small but effective steps towards awesomeness.

Most of us always wonder how machines can learn from data, and predict future based on the available information considering facts and scenarios. Today we are living in an era where most of us are working globally on big data technologies having great efficiency and speed. But having a huge amount of data only is not the complete solution and optimal use of data until we are able to find patterns out of it and use those patterns to predict future events and identify our interest specific solutions.

To understand how a machine can learn from past experiences and predict future based on the experience gained, we need to understand the working of a human brain first. Once we find out the pattern of a human brain to solve a problem, we can make our machine to learn in an almost same way. Almost same way because the human brain has no limits and there is a lot to explore. As machine learning is a huge field of study, and there are a lot of possibilities, here we are going to discuss one of the most simple algorithms of machine learning which is called Find-S algorithm.

What is Learning?

There are several definitions available on the internet of learning. One of the simplest definitions is “The activity or process of gaining knowledge or skill by studying, practicing, being taught, or experiencing something”. Similar to various definitions available of learning, there are various categories of learning methods.

As a human, we learn a lot of things during entire life. Some of them are based on our experience and some of them are based on memorization. On the basis of that we can divide learning methods into five parts:

1. Rote Learning (memorization): Memorizing things without knowing the concept/ logic behind them 2. Passive Learning (instructions): Learning from a teacher/expert. 3. Analogy (experience): Learning new things from our past experience. 4. Inductive Learning (experience): On the basis of past experience formulating a generalized concept . 5. Deductive Learning: Deriving new facts from past facts.

The inductive learning is based on formulating a generalized concept after observing a number of instances of examples of the concept . For an example, if a kid is asked to write an answer of 2*8=?. He/She can either use the rote learning method to memorize the answer, or he/she can use inductive learning using the examples like 2*1=2, 2*2=4… and so on, to formulate a concept to calculate the results. In this way, the kid will be able to solve a similar type of questions using the same concept.

Similarly, we can make our machine to learn from past data and make them intelligent to identify whether an object falls into a specific category of our interest or not.

What is concept learning?

In terms of machine learning, the concept learning can be formulated as “Problem of searching through a predefined space of potential hypotheses for the hypothesis that best fits the training examples” -Tom Michell .

Much of human learning involves acquiring general concepts from past experiences. For an example, humans identify different vehicles among all the vehicles based on some specific set of features defined over a large set of features. This special set of features differentiates the subset of cars in the set of vehicles. This set of features that differentiate cars, can be called a concept .

Similarly, machines can also learn from the concepts to identify whether an object belongs to a specific category or not by processing past/training data to find a hypothesis that best fits the training examples.

Target Concept:

The set of items/objects over which the concept is defined is called the set of instances and denoted by X. The concept or function to be learned is called the target concept and denoted by c. It can be seen as a boolean valued function defined over X and can be represented as:

c: X -> {0, 1}

So, if we have a set of training examples with specific features, of target concept c, the problem faced by the learner to estimate c that can be defined on training data. H is used to denote the set of all possible hypotheses that the learner may consider regarding the identity of the target concept. The goal of a learner is to find a hypothesis h which can identify all the objects in X so that:

h(x) = c(x) for all x in X

In this way there are three necessary things for an algorithm which supports concept learning:

1. Training data (Past experiences to train our models) 2. Target Concept (Hypothesis to identify data objects) 3. Actual data objects (For testing the models)

Inductive Learning Hypothesis:

As we discussed earlier, the ultimate goal of concept learning is to identify a hypothesis ‘h’ identical to target concept c over data set X with only available information about c is its value over X. So our algorithm can guaranty that it best fits on training data. In other words “ Any hypothesis found approximate the target function well over a sufficiently large set of training examples will also approximate the target function well over other unobserved examples “ .

Considering an example if a person goes to a movie or not based on 4 binary features with 2 values (true or false) possible:

1. Has Money -> <true, false> 2. Has Free Time -> <true, false> 3. It’s a Holiday -> <true, false> 4. Has Pending work -> <true, false>

And training data we have with two data objects as positive samples and one as negative:

x1 : <true, true, false, false> : +ve x2 : <true, false, false, true> : +ve x3: <true, false, false, true> : -ve

Hypothesis Notations: Each of the data objects represents a concept and hypotheses. Considering a hypothesis <true, true, false, false> is more specific because it can cover only one sample. To make a more generalized concept we can add some notations into this hypothesis. To this task we have following notations:

1. ⵁ (Represents a hypothesis which rejects all.) 2. < ? , ? , ? , ? > (Accepts all) 3. <true, false, ? , ? > (Accepts some)

The hypothesis ⵁ will reject all the data samples. The hypothesis <? , ? , ? , ? > will accept all the data samples. The ‘?’ notation indicates that the values of this specific feature do not affect the result.

In this way the total number of the possible hypothesis are: (3 * 3 * 3 * 3) + 1 where 3 because one feature can have either true, false or ‘?’ and one hypothesis for rejects all (ⵁ).

General to the specific ordering of hypothesis:

Many machine learning algorithms rely on the concept general-to-specific ordering of hypothesis.

h1 = < true, true, ?, ? > h2 = < true, ? , ? , ? >

Any instance classified by h1 will also be classified by h2. So we can say that h2 is more general than h1. Using this concept we can find a general hypothesis that can be defined over entire data set X.

Find-S Algorithm: Finding Maximally Specific Hypothesis:

To find out a single hypothesis defined on X we can use the concept more-general-then partial ordering. One way to do this is start with the most specific hypothesis from H and generalize this hypothesis each time it fails to classify and observed positive training data object as positive.

Step 1. The first step in Find-S algorithm is to start with most specific hypothesis that can be denoted by

h <- <ⵁ, ⵁ, ⵁ, ⵁ>

Step 2. This step involves picking up next training sample and apply step 3 on the sample.

Step 3. The next step involves observing the data sample. If the sample is negative the hypothesis remains unchanged and we pick next training sample by processing step 2 again else we process step 4.

Step 4. If the sample is positive and we find that our initial hypothesis is too specific because it does not cover the current training sample then we need to update our current hypothesis. This can be done by the pairwise conjunction (Logical AND operation) of current hypothesis and training sample.

If next training sample is <true, true, false, false> and current hypothesis is <ⵁ, ⵁ, ⵁ, ⵁ>, then we can directly replace our existing hypothesis with the new one.

If the next positive training sample is <true, true, false, true> and current hypothesis is <true, true, false, false> then we can perform a pairwise conjunctive AND with the current hypothesis and next training sample and find new hypothesis by putting ‘?’ in the place where the result of conjunction is false: <true, true, false, true> ⴷ <true, true, false, false> = <true, true, false, ?> Now we can replace our existing hypothesis with the new one: h <-<true, true, false, ?>

Step 5. This step involves repetition of step 2 till we have more training samples.

Step 6. Once there are no training samples the current hypothesis is the one we were trying to find. We can use the final hypothesis for classifying the real objects.

A concise form of the Find-S algorithm:

Step 1. Start with h = ⵁ Step 2. Use next input {x, c(x)} Step 3. If c(x) = 0, go to step 2 Step 4. h <- h ⴷ x (Pairwise AND) Step 5. If more examples : Go to step 2 Step 6. Stop

Limitations of the Find-S algorithm:

Find-S algorithm for concept learning is one of the most basic algorithms of machine learning with some limitation and disadvantages. Some of them are listed here:

1. No way to determine if the only final hypothesis (found by Find-S) is consistent with data or there are more hypothesis that is consistent with data.

2. Inconsistent sets of training examples can mislead the finds algorithm as it ignores negative data samples, so an algorithm that can detect inconsistency of training data would be better to use.

3. A good concept learning algorithm should be able to backtrack the choice of hypothesis found so that the resulting hypothesis can be improved over time. Unfortunately, Find-S provide no such method.

Many of the limitations can be removed in one most important algorithm of concept learning called Candidate elimination algorithm.

In our next blog, we will be explaining Find-S algorithm with a basic example. To explore Find-S implementation please visit the next part of this blog(Coming Soon).

Reference: Machine Learning-Tom-Michell

Share the Knol:

Click to share on LinkedIn (Opens in new window)
Click to share on Twitter (Opens in new window)
Click to share on Telegram (Opens in new window)
Click to share on Facebook (Opens in new window)

Written by Girish Bharti

Girish is a Software Consultant, with experience of more than 3.5 years. He is a scala developer and very passionate about his interest towards Scala Eco-system. He has also done many projects in different languages like Java and Asp.net. He can work in both supervised and unsupervised environment and have a craze for computers whether working or not, he is almost always in front of his laptop's screen. His hobbies include reading books and listening to music. He is self motivated, dedicated and focused towards his work. He believes in developing quality products. He wants to work on different projects and different domains. He is curious to gain knowledge of different domains and try to provide solutions that can utilize resources and improve performance. His personal interests include reading books, video games, cricket and social networking. He has done Masters in Computer Applications from Lal Bahadur Shastri Institute of Management, New Delhi.

6 thoughts on “ Concept Learning: The stepping stone towards Machine Learning with Find-S 9 min read ”

Pingback: First step Towards Machine Learning | Knoldus
Pingback: Concept Learning: Find-S implementation with Scala | Knoldus
Pingback: Artificial Intelligence vs Machine Learning vs Deep Learning | Anuj's Blog
Pingback: Artificial Intelligence vs Machine Learning vs Deep Learning | Knoldus

Reblogged this on Coding, Unix & Other Hackeresque Things .

Pingback: Artificial Intelligence Vs Machine Learning Vs Deep Learning – All About Artificial Intelligence

Comments are closed.

Discover more from Knoldus Blogs

Subscribe now to keep reading and get access to the full archive.

Type your email…

Chapter 2: Concept Learning and the General-to-Specific Ordering

Concept Learning: Inferring a boolean valued function from training examples of its input and output.
X: set of instances
x: one instance
c: target concept, c:X → {0, 1}
< x, c(x) >, training instance, can be a positive example or a negative example
D: set of training instances
H: set of possible hypotheses
h: one hypothesis, h: X → { 0, 1 }, the goal is to find h such that h(x) = c(x) for all x in X

Inductive Learning Hypothesis

Any hypothesis found to approximate the target function well over a sufficiently large set of training examples will also approximate the target function well over other unobserved examples.

Let h j and h k be boolean-valued functions defined over X. h j is more general than or equal to h k (written h j ≥ g h k ) if and only if (∀ x ∈ X) [ (h k (x) = 1) → (h j (x) = 1)]

This is a partial order since it is reflexive, antisymmetric and transitive.

Find-S Algorithm

Outputs a description of the most specific hypothesis consistent with the training examples.

Initialize h to the most specific hypothesis in H
If the constraint a i is NOT satisfied by x, then replace a i in h by the next more general constraint that is satisfied by x.
Output hypothesis h

For this particular algorithm, there is a bias that the target concept can be represented by a conjunction of attribute constraints.

Candidate Elimination Algorithm

Outputs a description of the set of all hypotheses consistent with the training examples.

A hypothesis h is consistent with a set of training examples D if and only if h(x) = c(x) for each example < x, c(x) > in D. Consistent(h, D) ≡ (∀ < x, c(x) > ∈ D) h(x) = c(x)

The version space denoted VS H,D with respect to hypothesis space H and training examples D, is the subset of hypotheses from H consistent with the training examples in D. VS H,D ≡ { h ∈ H | Consistent(h, D) }

The general boundary G, with respect to hypothesis space H and training data D, is the set of maximally general members of H consistent with D.

The specific boundary S, with respect to hypothesis space H and training data D, is the set of maximally specific members of H consistent with D.

Version Space Representation

Let X be an arbitrary set of instances and let H be a set of boolean-valued hypotheses defined over X. Let c:X → {0,1} be an arbitrary target concept defined over X, and let D be an arbitrary set of training examples {<x, c(x)>}. For all X, H, c and D such that S and G are well defined, VS H,D = {h ∈ H | (∃s ∈ S) (∃g ∈ G) (g ≥ g h ≥ g s)}

Initialize G to the set of maximally general hypotheses in H
Initialize S to the set of maximally specific hypotheses in H
Remove from G any hypothesis inconsistent with d
Remove s from S
Add to S all minimal generalizations h of s such that h is consistent with d, and some member of G is more general than h
Remove from S any hypothesis that is more general than another hypothesis in S
Remove from S any hypothesis inconsistent with d
Remove g from G
Add to G all minimal specializations h of g such that h is consistent with d, and some member of S is more specific than h
Remove from G any hypothesis that is less general than another hypothesis in G

Candidate Elimination Algorithm Issues

Will it converge to the correct hypothesis? Yes, if (1) the training examples are error free and (2) the correct hypothesis can be represented by a conjunction of attributes.
If the learner can request a specific training example, which one should it select?
How can a partially learned concept be used?

Inductive Bias

Definition: Consider a concept learning algorithm L for the set of instances X. Let c be an arbitrary concept defined over X and let D c = {<x, c(x)>} be an arbitrary set of training examples of c. Let L(x i , D c ) denote the classification assigned to the instance x i by L after training on the data D c . The inductive bias of L is any minimal set of assertions B such that for any target concept c and corresponding training examples D c (∀ x i ∈ X) [ L(x i , D c ) follows deductively from (B ∧ D c ∧ x i ) ]
Thus, one advantage of an inductive bias is that it gives the learner a rational basis for classifying unseen instances.
What is another advantage of bias?
What is one disadvantage of bias?
What is the inductive bias of the candidate elimination algorithm? Answer: the target concept c is a conjunction of attributes.
What is meant by a weak bias versus a strong bias?

Sample Exercise

Work exercise 2.4 on page 48.

General-To-Specific Ordering of Hypothesis

The theories can be sorted from the most specific to the most general. This will allow the machine learning algorithm to thoroughly investigate the hypothesis space without having to enumerate each and every hypothesis in it, which is impossible when the hypothesis space is infinitely vast.

Now we’ll speak about general-to-specific ordering and how to utilize it to construct a feeling of order in a hypothesis space in any concept learning issue.

In this article, we’ll have a look at the general-to-specific ordering of hypotheses.

Let us have a look at our previous EnjoySport example again,

Task T: Determine the value of EnjoySport for every given day based on the values of the day’s qualities.

The total proportion of days (EnjoySport) accurately anticipated is the performance metric P.

Experience E: A collection of days with pre-determined labels (EnjoySport: Yes/No).

Each hypothesis can be considered as a set of six constraints, with the values of the six attributes Sky, AirTemp, Humidity, Wind, Water, and Forecast specified.

Sky	Air temp	Humidity	Wind	Water	Forecast	EnjoySport
Sunny	Warm	Normal	Strong	Warm	Same	Yes
Sunny	Warm	High	Strong	Warm	Same	Yes
Rainy	Cold	High	Strong	Warm	Change	No
Sunny	Warm	High	Strong	Cool	Change	Yes

Take a look at the following two hypotheses:

h1 = <Rainy, Warm, Strong>

h2 = <Rainy, ?, Strong>

The question is how many and which examples are classed as positive by each of these theories (i.e., satisfy these hypotheses). Only example 4 is satisfactory for h1, however, both examples 3 and 4 are satisfactory and categorized as positive for h2.

What is the reason behind this? What makes these two hypotheses so different? The solution is found in the rigor with which each of these theories imposes limits. As you can see, h1 places more restrictions on you than h2! Naturally, h2 can categorize more good cases than h1! In this case, we may really assert the following:

“If an example meets h1, it will almost certainly meet h2, but not the other way around.”

This is due to the fact that h2 is more general than h1. This may be seen in the following example: h2 has a wider range of choices than h1. If an instance has the following values:< Rainy, Freezing, Strong>, h2 will classify it as positive, but h1 will not be fulfilled.

However, if h1 identifies an occurrence as positive, such as <Rainy, Warm, Strong>, h2 will almost certainly categorise it as positive as well.

In fact, each case that is categorised as positive by h1 is likewise classed as positive by h2. As a result, we might conclude that h2 is more generic than h1.

We state that x fulfils h if and only if h(x) = 1 for each instance x in X and hypothesis h in H.

Definition:

Let hj and hk be boolean-valued functions that are defined over X. If and only if, hj is more general than or equal to hk (written hj >=g hk).

We can show this relationship with the following notation:

The letter g stands for “general.” There are times when one theory is more general than the other, but it is not the same.

Because every case that fulfils hl also satisfies h2, hypothesis h2 is more general than hl.

In the same way, h2 is a more broad term than h3.

It’s worth noting that neither hl nor h3 are more general than the other; while the instances met by both hypotheses overlap, neither set subsumes the other.

A handful of the key algorithms that may be used to explore the hypothesis space, H, by making use of the g operation. Finding-S is the name of the method, with S standing for specific and implying that the purpose is to identify the most particular hypothesis.

We can observe that all the occurrences that fulfill both h1 and h3 also satisfy h2, thus we can conclude that:

h2 ≥g. h1 and h3 are two different types of h2 g h1 and h3.

FIND S Algorithm in Python

Computer graphics opengl mini projects, download final year projects, python program to implement find s algorithm – to get maximally specific hypothesis.

Exp. No. 1. Implement and demonstrate the FIND-S algorithm in Python for finding the most specific hypothesis based on a given set of training data samples. Read the training data from a .CSV file

Find-S Algorithm Machine Learning

Python program to implement and demonstrate find-s algorithm.

EnjoySport Dataset is saved as .csv (comma separated values) file in the current working directory otherwise use the complete path of the dataset set in the program:

sky	airtemp	humidity	wind	water	forcast	enjoysport
sunny	warm	normal	strong	warm	same	yes
sunny	warm	high	strong	warm	same	yes
rainy	cold	high	strong	warm	change	no
sunny	warm	high	strong	cool	change	yes

[[‘sky’, ‘airtemp’, ‘humidity’, ‘wind’, ‘water’, ‘forcast’, ‘enjoysport’], [‘sunny’, ‘warm’, ‘normal’, ‘strong’, ‘warm’, ‘same’, ‘yes’], [‘sunny’, ‘warm’, ‘high’, ‘strong’, ‘warm’, ‘same’, ‘yes’], [‘rainy’, ‘cold’, ‘high’, ‘strong’, ‘warm’, ‘change’, ‘no’], [‘sunny’, ‘warm’, ‘high’, ‘strong’, ‘cool’, ‘change’, ‘yes’]]

The total number of training instances are : 5

The initial hypothesis is : [‘0’, ‘0’, ‘0’, ‘0’, ‘0’, ‘0’]

Instance 2 is [‘sunny’, ‘warm’, ‘normal’, ‘strong’, ‘warm’, ‘same’, ‘yes’] and is Positive Instance

The hypothesis for the training instance 2 is: [‘sunny’, ‘warm’, ‘normal’, ‘strong’, ‘warm’, ‘same’]

Instance 3 is [‘sunny’, ‘warm’, ‘high’, ‘strong’, ‘warm’, ‘same’, ‘yes’] and is Positive Instance

The hypothesis for the training instance 3 is: [‘sunny’, ‘warm’, ‘?’, ‘strong’, ‘warm’, ‘same’]

Instance 4 is [‘rainy’, ‘cold’, ‘high’, ‘strong’, ‘warm’, ‘change’, ‘no’] and is Negative Instance Hence Ignored

The hypothesis for the training instance 4 is: [‘sunny’, ‘warm’, ‘?’, ‘strong’, ‘warm’, ‘same’]

Instance 5 is [‘sunny’, ‘warm’, ‘high’, ‘strong’, ‘cool’, ‘change’, ‘yes’] and is Positive Instance

The hypothesis for the training instance 5 is: [‘sunny’, ‘warm’, ‘?’, ‘strong’, ‘?’, ‘?’]

The Maximally specific hypothesis for the training instance is [‘sunny’, ‘warm’, ‘?’, ‘strong’, ‘?’, ‘?’]

Solved Numerical Example – Find S Algorithm to Find the Most Specific Hypothesis

This tutorial discusses how to Implement and demonstrate the FIND-S algorithm in Python for finding the most specific hypothesis based on a given set of training data samples. The training data is read from a .CSV file. If you like the tutorial share with your friends. Like the Facebook page for regular updates and YouTube channel for video tutorials.

Welcome to VTUPulse.com

Computer graphics and image processing mini projects -> click here, download final year project -> click here.

This will close in 12 seconds

Privacy Policy
Machine Learning Lab

TUTORIALTPOINT- Java Tutorial, C Tutorial, DBMS Tutorial

Computer Fundamentals

C Programming

C++ Programming
Java Programming
Python Programming

Machine Learning

Artificial Intelligence
Probability and statistics

Data Base Management

Data Science
Computer Organization
Cloud Computing
Operating System

Data Structures

C++ Programs
Java Programs
Data Structures Python
Python Programs

Saturday 16 July 2022

Find-s algorithm: finding maximally specific hypotheses.

Now after learning the concept of general-to-specific ordering of hypotheses, Now its time to use this partial ordering to organize the search for a hypothesis, that is consistent with the observed training examples. One way is to begin with the most specific possible hypothesis in H, then generalize this hypothesis each time it fails to cover an observed positive training example. FIND-S algorithm is used for this purpose. Here are the steps for find-s algorithm.

To illustrate this algorithm, assume the learner is given the sequence of training examples from the EnjoySport task

The first step of FIND-S is to initialize h to the most specific hypothesis in H h — (Ø, Ø, Ø, Ø, Ø, Ø)
First training example x1 = < Sunny, Warm, Normal, Strong ,Warm ,Same>, EnjoySport = +ve. Observing the first training example, it is clear that hypothesis h is too specific. None of the “Ø” constraints in h are satisfied by this example, so each is replaced by the next more general constraint that fits the example h1 = < Sunny, Warm, Normal, Strong ,Warm, Same>.
Consider the second training example x2 = < Sunny, Warm, High, Strong, Warm, Same>, EnjoySport = +ve. The second training example forces the algorithm to further generalize h, this time substituting a “?” in place of any attribute value in h that is not satisfied by the new example. Now h2 =< Sunny, Warm, ?, Strong, Warm, Same>
Consider the third training example x3 =< Rainy, Cold, High, Strong, Warm, Change>,EnjoySport = — ve. The FIND-S algorithm simply ignores every negative example. So the hypothesis remain as before, so h3 =< Sunny, Warm, ?, Strong, Warm, Same>
Consider the fourth training example x4 =<Sunny,Warm,High,Strong, Cool,Change>, EnjoySport =+ve. The fourth example leads to a further generalization of h as h4 =< Sunny, Warm, ?, Strong, ?, ?>
So the final hypothesis is < Sunny, Warm, ?, Strong, ?, ?>

The search begins (ho) with the most specific hypothesis in H, then considers increasingly general hypotheses (hl through h4) as mandated by the training examples. The search moves from hypothesis to hypothesis, searching from the most specific to progressively more general hypotheses along one chain of the partial ordering. At each step, the hypothesis is generalized only as far as necessary to cover the new positive example. Therefore, at each stage the hypothesis is the most specific hypothesis consistent with the training examples observed up to this point.

The key property of the FIND-S algorithm —

FIND-S is guaranteed to output the most specific hypothesis within H that is consistent with the positive training examples
FIND-S algorithm’s final hypothesis will also be consistent with the negative examples provided the correct target concept is contained in H, and provided the training examples are correct.

Unanswered questions by FIND-S

Has the learner converged to the correct target concept ?. Although FIND-S will find a hypothesis consistent with the training data, it has no way to determine whether it has found the only hypothesis in H consistent with the data (i.e., the correct target concept), or whether there are many other consistent hypotheses as well.
Why prefer the most specific hypothesis ?. In case there are multiple hypotheses consistent with the training examples, FIND-S will find the most specific. It is unclear whether we should prefer this hypothesis over, say, the most general, or some other hypothesis of intermediate generality.
Are the training examples consistent ?. In most practical learning problems there is some chance that the training examples will contain at least some errors or noise. Such inconsistent sets of training examples can severely mislead FIND-S, given the fact that it ignores negative examples.
What if there are several maximally specific consistent hypotheses?. There can be several maximally specific hypotheses consistent with the data. Find S finds only one

0 comments :

ML – Candidate Elimination Algorithm

The candidate elimination algorithm incrementally builds the version space given a hypothesis space H and a set E of examples. The examples are added one by one; each example possibly shrinks the version space by removing the hypotheses that are inconsistent with the example. The candidate elimination algorithm does this by updating the general and specific boundary for each new example.

You can consider this as an extended form of the Find-S algorithm.
Consider both positive and negative examples.
Actually, positive examples are used here as the Find-S algorithm (Basically they are generalizing from the specification).
While the negative example is specified in the generalizing form.

Terms Used:

Concept learning: Concept learning is basically the learning task of the machine (Learn by Train data)
General Hypothesis: Not Specifying features to learn the machine.
G = {‘?’, ‘?’,’?’,’?’…}: Number of attributes
Specific Hypothesis: Specifying features to learn machine (Specific feature)
S= {‘pi’,’pi’,’pi’…}: The number of pi depends on a number of attributes.
Version Space: It is an intermediate of general hypothesis and Specific hypothesis. It not only just writes one hypothesis but a set of all possible hypotheses based on training data-set.

Consider the dataset given below:

Algorithmic steps:

The Candidate Elimination Algorithm (CEA) is an improvement over the Find-S algorithm for classification tasks. While CEA shares some similarities with Find-S, it also has some essential differences that offer advantages and disadvantages. Here are some advantages and disadvantages of CEA in comparison with Find-S:

Advantages of CEA over Find-S:

Improved accuracy: CEA considers both positive and negative examples to generate the hypothesis, which can result in higher accuracy when dealing with noisy or incomplete data.
Flexibility: CEA can handle more complex classification tasks, such as those with multiple classes or non-linear decision boundaries.
More efficient: CEA reduces the number of hypotheses by generating a set of general hypotheses and then eliminating them one by one. This can result in faster processing and improved efficiency.
Better handling of continuous attributes: CEA can handle continuous attributes by creating boundaries for each attribute, which makes it more suitable for a wider range of datasets.

Disadvantages of CEA in comparison with Find-S:

More complex: CEA is a more complex algorithm than Find-S, which may make it more difficult for beginners or those without a strong background in machine learning to use and understand.
Higher memory requirements: CEA requires more memory to store the set of hypotheses and boundaries, which may make it less suitable for memory-constrained environments.
Slower processing for large datasets: CEA may become slower for larger datasets due to the increased number of hypotheses generated.
Higher potential for overfitting: The increased complexity of CEA may make it more prone to overfitting on the training data, especially if the dataset is small or has a high degree of noise.

Please Login to comment...

Improve your Coding Skills with Practice

What kind of Experience do you want to share?

Stack Overflow for Teams Where developers & technologists share private knowledge with coworkers
Advertising & Talent Reach devs & technologists worldwide about your product, service or employer brand
OverflowAI GenAI features for Teams
OverflowAPI Train & fine-tune LLMs
Labs The future of collective knowledge sharing
About the company Visit the blog

Collectives™ on Stack Overflow

Find centralized, trusted content and collaborate around the technologies you use most.

Q&A for work

Connect and share knowledge within a single location that is structured and easy to search.

Get early access and see previews of new features.

Is there any difference between most specific hypotheses obtained by Candidate Elimination and Find-S methods?

In terms of machine learning. Is there any difference between most specific hypotheses obtained by Candidate Elimination and Find-S methods?

Many Thanks

machine-learning

If there are several maximally specific hypotheses that fit a data set, Find-S will just return one of them, where as C-E will return all of them as part of the specific boundary of the version space.

If there is only 1 maximally specific hypothesis though, there is no difference.

Hope this helps!

Your Answer

Reminder: Answers generated by artificial intelligence tools are not allowed on Stack Overflow. Learn more

Sign up or log in

Post as a guest.

Required, but never shown

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy .

Not the answer you're looking for? Browse other questions tagged machine-learning or ask your own question .

The Overflow Blog
At scale, anything that could fail definitely will
Featured on Meta
Announcing a change to the data-dump process
Bringing clarity to status tag usage on meta sites
What does a new user need in a homepage experience on Stack Overflow?
Feedback requested: How do you use tag hover descriptions for curating and do...
Staging Ground Reviewer Motivation

Hot Network Questions

How to find the x-coordinate of the point circled(non-differentiable points) in this trigonometric function graph?
Fill the grid with numbers to make all four equations true
How can coordinates be meaningless in General Relativity?
Is consciousness a prerequisite for knowledge?
Second Derivative for Non standard Calculus
Do I need to validate a Genoa MET daily ticket every time?
How do Trinitarian Christians defend the unfalsifiability of the Trinity?
When trying to find the quartiles for discrete data, do we round to the nearest whole number?
Hardware debouncing of 3.3V high signal for an ESP32 turned on via optocoupler
Do eternal ordinances such as the festival of unleavened bread pose a biblical contradiction?
How would humans actually colonize mars?
Can my employer require me to wear a dirty uniform and refuse to provide additional shirts?
How would you slow the speed of a rogue solar system?
Hip pain when cycling (experienced cyclist)
Why is a USB memory stick getting hotter when connected to USB-3 (compared to USB-2)?
Lore reasons for being faithless
Maximizing the common value of both sides of an equation (part 2)
Nearly stalled on takeoff after just 3 hours training on a PPL. Is this normal?
Not a cross, not a word (number crossword)
World Building Knowledgebase - How to write good Military World Building
Admissibility of withdrawn confession
If a Palestinian converts to Judaism, can they get Israeli citizenship?
Why are poverty definitions not based off a person's access to necessities rather than a fixed number?
MANIFEST_UNKNOWN error: OCI index found, but Accept header does not support OCI indexes

Machine Learning

Artificial Intelligence

Control System

Supervised Learning

Classification, miscellaneous, related tutorials.

Interview Questions

The hypothesis is a common term in Machine Learning and data science projects. As we know, machine learning is one of the most powerful technologies across the world, which helps us to predict results based on past experiences. Moreover, data scientists and ML professionals conduct experiments that aim to solve a problem. These ML professionals and data scientists make an initial assumption for the solution of the problem.

This assumption in Machine learning is known as Hypothesis. In Machine Learning, at various times, Hypothesis and Model are used interchangeably. However, a Hypothesis is an assumption made by scientists, whereas a model is a mathematical representation that is used to test the hypothesis. In this topic, "Hypothesis in Machine Learning," we will discuss a few important concepts related to a hypothesis in machine learning and their importance. So, let's start with a quick introduction to Hypothesis.

It is just a guess based on some known facts but has not yet been proven. A good hypothesis is testable, which results in either true or false.

: Let's understand the hypothesis with a common example. Some scientist claims that ultraviolet (UV) light can damage the eyes then it may also cause blindness.

In this example, a scientist just claims that UV rays are harmful to the eyes, but we assume they may cause blindness. However, it may or may not be possible. Hence, these types of assumptions are called a hypothesis.

The hypothesis is one of the commonly used concepts of statistics in Machine Learning. It is specifically used in Supervised Machine learning, where an ML model learns a function that best maps the input to corresponding outputs with the help of an available dataset.

There are some common methods given to find out the possible hypothesis from the Hypothesis space, where hypothesis space is represented by and hypothesis by Th ese are defined as follows:

It is used by supervised machine learning algorithms to determine the best possible hypothesis to describe the target function or best maps input to output.

It is often constrained by choice of the framing of the problem, the choice of model, and the choice of model configuration.

. It is primarily based on data as well as bias and restrictions applied to data.

Hence hypothesis (h) can be concluded as a single hypothesis that maps input to proper output and can be evaluated as well as used to make predictions.

The hypothesis (h) can be formulated in machine learning as follows:

Where,

Y: Range

m: Slope of the line which divided test data or changes in y divided by change in x.

x: domain

c: intercept (constant)

: Let's understand the hypothesis (h) and hypothesis space (H) with a two-dimensional coordinate plane showing the distribution of data as follows:

Hypothesis space (H) is the composition of all legal best possible ways to divide the coordinate plane so that it best maps input to proper output.

Further, each individual best possible way is called a hypothesis (h). Hence, the hypothesis and hypothesis space would be like this:

Similar to the hypothesis in machine learning, it is also considered an assumption of the output. However, it is falsifiable, which means it can be failed in the presence of sufficient evidence.

Unlike machine learning, we cannot accept any hypothesis in statistics because it is just an imaginary result and based on probability. Before start working on an experiment, we must be aware of two important types of hypotheses as follows:

A null hypothesis is a type of statistical hypothesis which tells that there is no statistically significant effect exists in the given set of observations. It is also known as conjecture and is used in quantitative analysis to test theories about markets, investment, and finance to decide whether an idea is true or false. An alternative hypothesis is a direct contradiction of the null hypothesis, which means if one of the two hypotheses is true, then the other must be false. In other words, an alternative hypothesis is a type of statistical hypothesis which tells that there is some significant effect that exists in the given set of observations.

The significance level is the primary thing that must be set before starting an experiment. It is useful to define the tolerance of error and the level at which effect can be considered significantly. During the testing process in an experiment, a 95% significance level is accepted, and the remaining 5% can be neglected. The significance level also tells the critical or threshold value. For e.g., in an experiment, if the significance level is set to 98%, then the critical value is 0.02%.

The p-value in statistics is defined as the evidence against a null hypothesis. In other words, P-value is the probability that a random chance generated the data or something else that is equal or rarer under the null hypothesis condition.

If the p-value is smaller, the evidence will be stronger, and vice-versa which means the null hypothesis can be rejected in testing. It is always represented in a decimal form, such as 0.035.

Whenever a statistical test is carried out on the population and sample to find out P-value, then it always depends upon the critical value. If the p-value is less than the critical value, then it shows the effect is significant, and the null hypothesis can be rejected. Further, if it is higher than the critical value, it shows that there is no significant effect and hence fails to reject the Null Hypothesis.

In the series of mapping instances of inputs to outputs in supervised machine learning, the hypothesis is a very useful concept that helps to approximate a target function in machine learning. It is available in all analytics domains and is also considered one of the important factors to check whether a change should be introduced or not. It covers the entire training data sets to efficiency as well as the performance of the models.

Hence, in this topic, we have covered various important concepts related to the hypothesis in machine learning and statistics and some important parameters such as p-value, significance level, etc., to understand hypothesis concepts in a better way.

Send your Feedback to [email protected]

Help Others, Please Share

Learn Latest Tutorials

Transact-SQL

Reinforcement Learning

R Programming

React Native

Python Design Patterns

Python Pillow

Python Turtle

Preparation

Verbal Ability

Company Questions

Trending Technologies

Cloud Computing

Data Science

B.Tech / MCA

Data Structures

Operating System

Computer Network

Compiler Design

Computer Organization

Discrete Mathematics

Ethical Hacking

Computer Graphics

Software Engineering

Web Technology

Cyber Security

C Programming

Data Mining

Data Warehouse

This week: the arXiv Accessibility Forum

Help | Advanced Search

Computer Science > Robotics

Title: flowretrieval: flow-guided data retrieval for few-shot imitation learning.

Abstract: Few-shot imitation learning relies on only a small amount of task-specific demonstrations to efficiently adapt a policy for a given downstream tasks. Retrieval-based methods come with a promise of retrieving relevant past experiences to augment this target data when learning policies. However, existing data retrieval methods fall under two extremes: they either rely on the existence of exact behaviors with visually similar scenes in the prior data, which is impractical to assume; or they retrieve based on semantic similarity of high-level language descriptions of the task, which might not be that informative about the shared low-level behaviors or motions across tasks that is often a more important factor for retrieving relevant data for policy learning. In this work, we investigate how we can leverage motion similarity in the vast amount of cross-task data to improve few-shot imitation learning of the target task. Our key insight is that motion-similar data carries rich information about the effects of actions and object interactions that can be leveraged during few-shot adaptation. We propose FlowRetrieval, an approach that leverages optical flow representations for both extracting similar motions to target tasks from prior data, and for guiding learning of a policy that can maximally benefit from such data. Our results show FlowRetrieval significantly outperforms prior methods across simulated and real-world domains, achieving on average 27% higher success rate than the best retrieval-based prior method. In the Pen-in-Cup task with a real Franka Emika robot, FlowRetrieval achieves 3.7x the performance of the baseline imitation learning technique that learns from all prior and target data. Website: this https URL

Subjects:	Robotics (cs.RO); Machine Learning (cs.LG)
Cite as:	[cs.RO]
	(or [cs.RO] for this version)
	Focus to learn more arXiv-issued DOI via DataCite

Submission history

Access paper:.

HTML (experimental)
Other Formats

References & Citations

Google Scholar
Semantic Scholar

BibTeX formatted citation

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .

Decoding emotions: unveiling the potential of facial landmarks

Published: 03 September 2024

Cite this article

Junhyeok Jang ORCID: orcid.org/0009-0009-5604-2019 1 &
Jongwan Kim ORCID: orcid.org/0000-0003-1316-1041 1

This study aimed to validate emotion decoding using a facial landmark-based prediction model trained on real-world data. We used facial landmark data and behavioral ratings measured while 29 participants (21 female, 22 male; age = 19–29, M = 22.37, SD = 2.25). viewed short video clips eliciting 10 different emotions including amusement, anger, awe, disgust, enthusiasm, fear, liking, sadness, surprise, and neutrality. Cross-participant classifications with support vector machine classifier were conducted to predict emotions across individuals. The results demonstrated promising predictive capabilities, achieving over 40% accuracy for behavioral ratings, facial landmarks, and their combination. Notably, classification based on combined features reached 75% accuracy, excelling in emotions like disgust, fear, and surprise. In summary, this study demonstrates the ability of facial landmark-based model to decode emotions accurately, particularly emphasizing negative and high-arousal emotions, and underscores the importance of supplementary tools in enhancing emotional response prediction.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save.

Get 10 units per month
Download Article/Chapter or eBook
1 Unit = 1 Article or 1 Chapter
Cancel anytime

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

Explore related subjects

Artificial Intelligence

Data availability

This study is based on Saganowski’s experimental data(2022). The data are available in GitHub at https://github.com/Emognition/Emognition-wearable-dataset-2020 . These data were derived from the following resources available in the public domain: https://www.nature.com/articles/s41597-022-01262-0 , doi: https://doi.org/10.1038/s41597-022-01262-0 .

Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The Self-Assessment Manikin and the semantic differential. Journal of Behavior Therapy and Experimental Psychiatry , 25 (1), 49–59. https://doi.org/10.1016/0005-7916(94)90063-9

Article PubMed Google Scholar

Codispoti, M., Surcinelli, P., & Baldaro, B. (2008). Watching emotional movies: Affective reactions and gender differences. International Journal of Psychophysiology , 69 (2), 90–95. https://doi.org/10.1016/j.ijpsycho.2008.03.004

Croy, I., Laqua, K., Suss, F., Joraschky, P., Ziemssen, T., & Hummel, T. (2013). The sensory channel of presentation alters subjective ratings and autonomic responses toward disgusting stimuli-blood pressure, heart rate and skin conductance in response to visual, auditory, haptic and olfactory presented disgusting stimuli. Frontiers in Human Neuroscience , 7 (SEP), 510. https://doi.org/10.3389/fnhum.2013.00510

Article PubMed PubMed Central Google Scholar

Day, D. V., Schleicher, D. J., Unckless, A. L., & Hiller, N. J. (2002). Self-monitoring personality at work: A meta-analytic investigation of construct validity. Journal of Applied Psychology , 87 (2), 390–401. https://doi.org/10.1037/0021-9010.87.2.390

Dellacherie, D., Roy, M., Hugueville, L., Peretz, I., & Samson, S. (2011). The effect of musical experience on emotional self-reports and psychophysiological responses to dissonance. Psychophysiology , 48 (3), 337–349. https://doi.org/10.1111/j.1469-8986.2010.01075.x

Ekman, P. (1999). Basic emotions. Handbook of cognition and emotion , 98 (45–60), 16.

Google Scholar

Fredrickson, B. L. (2001). The role of positive emotions in positive psychology. The broaden-and-build theory of positive emotions. American Psychologist , 56 (3), 218–226. https://doi.org/10.1037//0003-066x.56.3.218

Fridlund, A. J., & Cacioppo, J. T. (1986). Guidelines for human electromyographic research. Psychophysiology , 23 (5), 567–589. https://doi.org/10.1111/j.1469-8986.1986.tb00676.x

Gomez, P., & Danuser, B. (2007). Relationships between musical structure and psychophysiological measures of emotion. Emotion , 7 (2), 377–387. https://doi.org/10.1037/1528-3542.7.2.377

Gorkiewicz, T., Danielewski, K., Andraka, K., Kondrakiewicz, K., Meyza, K., Kaminski, J., & Knapska, E. (2023). Social buffering diminishes fear response but does not equal improved fear extinction. Cerebral Cortex , 33 (8), 5007–5024. https://doi.org/10.1093/cercor/bhac395

Gross, J. J. (1998). Antecedent- and response-focused emotion regulation: Divergent consequences for experience, expression, and physiology. Journal of Personality and Social Psychology , 74 (1), 224–237. https://doi.org/10.1037//0022-3514.74.1.224

Kim, J., Wang, J., Wedell, D. H., & Shinkareva, S. V. (2016). Identifying core affect in individuals from fMRI responses to dynamic naturalistic audiovisual stimuli. PloS one , 11 (9), e0161589. https://doi.org/10.1371/journal.pone.0161589

Leventhal, H., & Scherer, K. (1987). The relationship of emotion to cognition: A functional approach to a semantic controversy. Cognition and Emotion , 1 (1), 3–28. https://doi.org/10.1080/02699938708408361

Article Google Scholar

Lewenberg, Y., Bachrach, Y., Shankar, S., & Criminisi, A. (2016, March). Predicting personal traits from facial images using convolutional neural networks augmented with facial landmark information. In Proceedings of the AAAI conference on artificial intelligence (Vol. 30, No. 1). https://doi.org/10.1609/aaai.v30i1.9844

Martell, R. F., Lane, D. M., & Emrich, C. (1996). Male-female differences: A computer simulation. American Psychologist , 51 (2), 157–158. https://doi.org/10.1037/0003-066x.51.2.157

Mukhiddinov, M., Djuraev, O., Akhmedov, F., Mukhamadiyev, A., & Cho, J. (2023). Masked face emotion recognition based on facial landmarks and deep learning approaches for visually impaired people. Sensors (Basel) , 23 (3), 1080. https://doi.org/10.3390/s23031080

Okubo, M., Kobayashi, A., & Ishikawa, K. (2012). A fake smile thwarts cheater detection. Journal of Nonverbal Behavior , 36 (3), 217–225. https://doi.org/10.1007/s10919-012-0134-9

Porges, S. W. (2009). Reciprocal influences between body and brain in the perception and expression of affect. The healing power of emotion: Affective neuroscience, development & clinical practice , 27–54.

Russell, J. A., Weiss, A., & Mendelsohn, G. A. (1989). Affect grid - a single-item scale of pleasure and arousal. Journal of Personality and Social Psychology , 57 (3), 493–502. https://doi.org/10.1037/0022-3514.57.3.493

Saganowski, S., Komoszynska, J., Behnke, M., Perz, B., Kunc, D., Klich, B., Kaczmarek, L. D., & Kazienko, P. (2022). Emognition dataset: Emotion recognition with self-reports, facial expressions, and physiology using wearables. Sci Data , 9 (1), 158. https://doi.org/10.1038/s41597-022-01262-0

Schmidt, K. L., & Cohn, J. F. (2001). Human facial expressions as adaptations: Evolutionary questions in facial expression research. American Journal of Physical Anthropology: The Official Publication of the American Association of Physical Anthropologists, 116 (S33), 3–24. https://doi.org/10.1002/ajpa.20001

Sullivan, L. A., & Harnish, R. J. (1990). Body-image - differences between high and low self-monitoring males and females. Journal of Research in Personality , 24 (3), 291–302. https://doi.org/10.1016/0092-6566(90)90022-X

Suthaharan, S., & Suthaharan, S. (2016). Support vector machine. Machine learning models and algorithms for big data classification: thinking with examples for effective learning , 207–235.

Tie, Y., & Guan, L. (2013). Automatic landmark point detection and tracking for human facial expressions. EURASIP Journal on Image and Video Processing , 2013 (1), 8. https://doi.org/10.1186/1687-5281-2013-8

Toisoul, A., Kossaifi, J., Bulat, A., Tzimiropoulos, G., & Pantic, M. (2021). Estimation of continuous valence and arousal levels from faces in naturalistic conditions. Nature Machine Intelligence , 3 (1), 42–50. https://doi.org/10.1038/s42256-020-00280-0

Download references

The research received funding from the Brain Korea 21 fourth project of the Korea Research Foundation (Jeonbuk National University, Psychology Department no. 4199990714213).

Author information

Authors and affiliations.

Department of Psychology, Jeonbuk National University, Jeonju-si, 54896, South Korea

Junhyeok Jang & Jongwan Kim

You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Jongwan Kim .

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Jang, J., Kim, J. Decoding emotions: unveiling the potential of facial landmarks. Curr Psychol (2024). https://doi.org/10.1007/s12144-024-06449-9

Download citation

Accepted : 21 July 2024

Published : 03 September 2024

DOI : https://doi.org/10.1007/s12144-024-06449-9

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Face landmark
Emotional experience
Naturalistic stimuli
Support vector machine
Find a journal
Publish with us
Track your research

Information

Author Services

Initiatives

You are accessing a machine-readable page. In order to be human-readable, please install an RSS reader.

All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess .

Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.

Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.

Original Submission Date Received: .

Active Journals
Find a Journal
Proceedings Series
For Authors
For Reviewers
For Editors
For Librarians
For Publishers
For Societies
For Conference Organizers
Open Access Policy
Institutional Open Access Program
Special Issues Guidelines
Editorial Process
Research and Publication Ethics
Article Processing Charges
Testimonials
Preprints.org
SciProfiles
Encyclopedia

Article Menu

Subscribe SciFeed
Recommended Articles
Google Scholar
on Google Scholar
Table of Contents

Find support for a specific problem in the support section of our website.

Please let us know what you think of our products and services.

Visit our dedicated information section to learn more about MDPI.

JSmol Viewer

Systematic review and clinical insights: the role of the ketogenic diet in managing glioblastoma in cancer neuroscience.

1. Introduction

1.1. ketogenic diet, 1.2. metabolic changes of neoplastic cells, 1.3. theories regarding carcinogenesis: a focus on mitochondria and relation to kd, 1.4. mechanism, 1.5. ketogenic diet and epigenetic targets, 1.6. preclinical studies, 2. materials and methods.

Preclinical studies done on animals
Letters or responses to the editor
Review articles
Studies not reporting feasibility and survival
Original articles;
Studies with KD as an intervention alone or in combination with other conventional and nonconventional treatments for malignant glioma;
All published articles to date regardless of age, gender, ethnicity, and country study conducted.

3.1. Published Materials

3.2. ongoing trials, 4. discussion, 5. conclusions, author contributions, institutional review board statement, informed consent statement, data availability statement, conflicts of interest.

Diwanji, T.P.; Engelman, A.; Snider, J.W.; Mohindra, P. Epidemiology, diagnosis, and optimal management of glioma in adolescents and young adults. Adolesc. Health Med. Ther. 2017 , 8 , 99–113. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Lim-Fat, M.J.; Das, S. Unique molecular, clinical, and treatment aspects of gliomas in adolescents and young adults: A review. J. Neurosurg. 2023 , 139 , 1619–1627. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Eisenstat, D.D.; Pollack, I.F.; Demers, A.; Sapp, M.V.; Lambert, P.; Weisfeld-Adams, J.D.; Burger, P.C.; Gilles, F.; Davis, R.L.; Packer, R.; et al. Impact of tumor location and pathological discordance on survival of children with midline high-grade gliomas treated on Children’s Cancer Group high-grade glioma study CCG-945. J. Neurooncol. 2015 , 121 , 573–581. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Bavle, A.; Chintagumpala, M. Pediatric high-grade glioma: A review of biology, prognosis, and treatment. J. Radiat. Oncol. 2018 , 7 , 7–15. [ Google Scholar ] [ CrossRef ]
Sievert, A.J.; Fisher, M.J. Pediatric low-grade gliomas. J. Child. Neurol. 2009 , 24 , 1397–1408. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Youland, R.S.; Khwaja, S.S.; Schomas, D.A.; Keating, G.F.; Wetjen, N.M.; Laack, N.N. Prognostic factors and survival patterns in pediatric low-grade gliomas over 4 decades. J. Pediatr. Hematol. Oncol. 2013 , 35 , 197–205. [ Google Scholar ] [ CrossRef ]
Ris, M.D.; Beebe, D.W. Neurodevelopmental outcomes of children with low-grade gliomas. Dev. Disabil. Res. Rev. 2008 , 14 , 196–202. [ Google Scholar ] [ CrossRef ]
Rees, J.H. Low-grade gliomas in adults. Curr. Opin. Neurol. 2002 , 15 , 657–661. [ Google Scholar ] [ CrossRef ]
Grier, J.T.; Batchelor, T. Low-grade gliomas in adults. Oncologist 2006 , 11 , 681–693. [ Google Scholar ] [ CrossRef ]
Artzi, M.; Liberman, G.; Vaisman, N.; Bokstein, F.; Vitinshtein, F.; Aizenstein, O.; Ben Bashat, D. Changes in cerebral metabolism during ketogenic diet in patients with primary brain tumors: (1)H-MRS study. J. Neurooncol. 2017 , 132 , 267–275. [ Google Scholar ] [ CrossRef ]
Foppiani, A.; De Amicis, R.; Lessa, C.; Leone, A.; Ravella, S.; Ciusani, E.; Silvani, A.; Zuccoli, G.; Battezzati, A.; Lamperti, E.; et al. Isocaloric Ketogenic Diet in Adults with High-Grade Gliomas: A Prospective Metabolic Study. Nutr. Cancer 2021 , 73 , 1004–1014. [ Google Scholar ] [ CrossRef ]
Sargaco, B.; Oliveira, P.A.; Antunes, M.L.; Moreira, A.C. Effects of the Ketogenic Diet in the Treatment of Gliomas: A Systematic Review. Nutrients 2022 , 14 , 1007. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Tagliabue, A.; Armeno, M.; Berk, K.A.; Guglielmetti, M.; Ferraris, C.; Olieman, J.; van der Louw, E. Ketogenic diet for epilepsy and obesity: Is it the same? Nutr. Metab. Cardiovasc. Dis. 2024 , 34 , 581–589. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Winter, S.F.; Loebel, F.; Dietrich, J. Role of ketogenic metabolic therapy in malignant glioma: A systematic review. Crit. Rev. Oncol. Hematol. 2017 , 112 , 41–58. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Poff, A.M.; Ari, C.; Seyfried, T.N.; D’Agostino, D.P. The ketogenic diet and hyperbaric oxygen therapy prolong survival in mice with systemic metastatic cancer. PLoS ONE 2013 , 8 , e65522. [ Google Scholar ]
Strowd, R.E., 3rd; Grossman, S.A. The Role of Glucose Modulation and Dietary Supplementation in Patients With Central Nervous System Tumors. Curr. Treat. Options Oncol. 2015 , 16 , 36. [ Google Scholar ] [ CrossRef ]
Poff, A.; Koutnik, A.P.; Egan, K.M.; Sahebjam, S.; D’Agostino, D.; Kumar, N.B. Targeting the Warburg effect for cancer treatment: Ketogenic diets for management of glioma. Semin. Cancer Biol. 2019 , 56 , 135–148. [ Google Scholar ] [ CrossRef ]
Valerio, J.E.; Wolf, A.; Wu, X.; Santiago Rea, N.; Fernandez Gomez, M.; Borro, M.; Alvarez-Pinzon, A.M. Assessment of Gamma Knife Stereotactic Radiosurgery as an Adjuvant Therapy in First-Line Management of Newly Diagnosed Glioblastoma: Insights from Ten Years at a Neuroscience Center. Int. J. Transl. Med. 2024 , 4 , 298–308. [ Google Scholar ] [ CrossRef ]
Kalyanaraman, B. Teaching the basics of cancer metabolism: Developing antitumor strategies by exploiting the differences between normal and cancer cell metabolism. Redox Biol. 2017 , 12 , 833–842. [ Google Scholar ] [ CrossRef ]
Xiao, F.; Wang, C.; Yin, H.; Yu, J.; Chen, S.; Fang, J.; Guo, F. Leucine deprivation inhibits proliferation and induces apoptosis of human breast cancer cells via fatty acid synthase. Oncotarget 2016 , 7 , 63679–63689. [ Google Scholar ] [ CrossRef ]
Chinopoulos, C.; Seyfried, T.N. Mitochondrial Substrate-Level Phosphorylation as Energy Source for Glioblastoma: Review and Hypothesis. ASN Neuro 2018 , 10 , 1759091418818261. [ Google Scholar ] [ CrossRef ]
Boveri, T. Concerning the origin of malignant tumours by Theodor Boveri. Translated and annotated by Henry Harris. J. Cell Sci. 2008 , 121 (Suppl. S1), 1–84. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Warburg, O. On the origin of cancer cells. Science 1956 , 123 , 309–314. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Seyfried, T.N.; Flores, R.E.; Poff, A.M.; D’Agostino, D.P. Cancer as a metabolic disease: Implications for novel therapeutics. Carcinogenesis 2014 , 35 , 515–527. [ Google Scholar ] [ CrossRef ]
Cairns, J. The origin of human cancers. Nature 1981 , 289 , 353–357. [ Google Scholar ] [ CrossRef ]
Szent-Gyorgyi, A. The living state and cancer. Proc. Natl. Acad. Sci. USA 1977 , 74 , 2844–2847. [ Google Scholar ] [ CrossRef ]
Mukherjee, S. The Emperor of All Maladies: A Biography of Cancer ; Simon and Schuster: New York, NY, USA, 2010; 579p. [ Google Scholar ]
Seyfried, T.N.; Chinopoulos, C. Can the Mitochondrial Metabolic Theory Explain Better the Origin and Management of Cancer than Can the Somatic Mutation Theory? Metabolites 2021 , 11 , 572. [ Google Scholar ] [ CrossRef ]
Mukherjee, P.; Augur, Z.M.; Li, M.; Hill, C.; Greenwood, B.; Domin, M.A.; Kondakci, G.; Narain, N.R.; Kiebish, M.A.; Bronson, R.T.; et al. Therapeutic benefit of combining calorie-restricted ketogenic diet and glutamine targeting in late-stage experimental glioblastoma. Commun. Biol. 2019 , 2 , 200. [ Google Scholar ] [ CrossRef ]
Dal Bello, S.; Valdemarin, F.; Martinuzzi, D.; Filippi, F.; Gigli, G.L.; Valente, M. Ketogenic Diet in the Treatment of Gliomas and Glioblastomas. Nutrients 2022 , 14 , 3851. [ Google Scholar ] [ CrossRef ]
Ji, C.C.; Hu, Y.Y.; Cheng, G.; Liang, L.; Gao, B.; Ren, Y.P.; Liu, J.T.; Cao, X.L.; Zheng, M.H.; Li, S.Z.; et al. A ketogenic diet attenuates proliferation and stemness of glioma stem-like cells by altering metabolism resulting in increased ROS production. Int. J. Oncol. 2020 , 56 , 606–617. [ Google Scholar ] [ CrossRef ]
Bowers, L.W.; Rossi, E.L.; O’Flanagan, C.H.; deGraffenried, L.A.; Hursting, S.D. The Role of the Insulin/IGF System in Cancer: Lessons Learned from Clinical Trials and the Energy Balance-Cancer Link. Front. Endocrinol. 2015 , 6 , 77. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Cairns, J.; Ung, C.Y.; Da Rocha, E.L.; Zhang, C.; Correia, C.; Weinshilboum, R.; Wang, L.; Li, H. A network-based phenotype mapping approach to identify genes that modulate drug response phenotypes. Sci. Rep. 2016 , 6 , 37003. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Mukherjee, P.; Mulrooney, T.J.; Marsh, J.; Blair, D.; Chiles, T.C.; Seyfried, T.N. Differential effects of energy stress on AMPK phosphorylation and apoptosis in experimental brain tumor and normal brain. Mol. Cancer 2008 , 7 , 37. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Simone, B.A.; Champ, C.E.; Rosenberg, A.L.; Berger, A.C.; Monti, D.A.; Dicker, A.P.; Simone, N.L. Selectively starving cancer cells through dietary manipulation: Methods and clinical implications. Future Oncol. 2013 , 9 , 959–976. [ Google Scholar ] [ CrossRef ]
Pazmandi, J.; O’Neill, K.S.; Scheck, A.C.; Szlosarek, P.W.; Woolf, E.C.; Brooks, K.S.; Syed, N. Abstract 240: The ketogenic diet alters the expression of microRNAs that play key roles in tumor development. Cancer Res. 2015 , 75 (Suppl. S15), 240. [ Google Scholar ] [ CrossRef ]
Poff, A.M.; Ari, C.; Arnold, P.; Seyfried, T.N.; D’Agostino, D.P. Ketone supplementation decreases tumor cell viability and prolongs survival of mice with metastatic cancer. Int. J. Cancer 2014 , 135 , 1711–1720. [ Google Scholar ] [ CrossRef ]
Christofk, H.R.; Vander Heiden, M.G.; Harris, M.H.; Ramanathan, A.; Gerszten, R.E.; Wei, R.; Fleming, M.D.; Schreiber, S.L.; Cantley, L.C. The M2 splice isoform of pyruvate kinase is important for cancer metabolism and tumour growth. Nature 2008 , 452 , 230–233. [ Google Scholar ] [ CrossRef ]
Hitosugi, T.; Kang, S.; Vander Heiden, M.G.; Chung, T.W.; Elf, S.; Lythgoe, K.; Dong, S.; Lonial, S.; Wang, X.; Chen, G.Z.; et al. Tyrosine phosphorylation inhibits PKM2 to promote the Warburg effect and tumor growth. Sci. Signal 2009 , 2 , ra73. [ Google Scholar ] [ CrossRef ]
Luo, W.; Hu, H.; Chang, R.; Zhong, J.; Knabel, M.; O’Meally, R.; Cole, R.N.; Pandey, A.; Semenza, G.L. Pyruvate kinase M2 is a PHD3-stimulated coactivator for hypoxia-inducible factor 1. Cell 2011 , 145 , 732–744. [ Google Scholar ] [ CrossRef ]
Fu, X.; Chin, R.M.; Vergnes, L.; Hwang, H.; Deng, G.; Xing, Y.; Pai, M.Y.; Li, S.; Ta, L.; Fazlollahi, F.; et al. 2-Hydroxyglutarate Inhibits ATP Synthase and mTOR Signaling. Cell Metab. 2015 , 22 , 508–515. [ Google Scholar ] [ CrossRef ]
Woolf, E.C.; Curley, K.L.; Liu, Q.; Turner, G.H.; Charlton, J.A.; Preul, M.C.; Scheck, A.C. The Ketogenic Diet Alters the Hypoxic Response and Affects Expression of Proteins Associated with Angiogenesis, Invasive Potential and Vascular Permeability in a Mouse Glioma Model. PLoS ONE 2015 , 10 , e0130357. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Lussier, D.M.; Woolf, E.C.; Johnson, J.L.; Brooks, K.S.; Blattman, J.N.; Scheck, A.C. Enhanced immunity in a mouse model of malignant glioma is mediated by a therapeutic ketogenic diet. BMC Cancer 2016 , 16 , 310. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Boison, D. New insights into the mechanisms of the ketogenic diet. Curr. Opin. Neurol. 2017 , 30 , 187–192. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Mondal, D.; Shinde, S.; Paul, S.; Thakur, S.; Velu, G.S.; Tiwari, A.K.; Dixit, V.; Amit, A.; Vishvakarma, N.K.; Shukla, D. Diagnostic significance of dysregulated miRNAs in T-cell malignancies and their metabolic roles. Front. Oncol. 2023 , 13 , 1230273. [ Google Scholar ] [ CrossRef ]
Shea, A.; Harish, V.; Afzal, Z.; Chijioke, J.; Kedir, H.; Dusmatova, S.; Roy, A.; Ramalinga, M.; Harris, B.; Blancato, J.; et al. MicroRNAs in glioblastoma multiforme pathogenesis and therapeutics. Cancer Med. 2016 , 5 , 1917–1946. [ Google Scholar ] [ CrossRef ]
Zeng, Q.; Stylianou, T.; Preston, J.; Glover, S.; O’Neill, K.; Woolf, E.C.; Scheck, A.C.; Syed, N. The ketogenic diet alters the epigenetic landscape of GBM to potentiate the effects of chemotherapy and radiotherapy. Neuro-Oncology 2019 , 21 (Suppl. S4), iv8. [ Google Scholar ] [ CrossRef ]
Lv, M.; Zhu, X.; Wang, H.; Wang, F.; Guan, W. Roles of caloric restriction, ketogenic diet and intermittent fasting during initiation, progression and metastasis of cancer in animal models: A systematic review and meta-analysis. PLoS ONE 2014 , 9 , e115147. [ Google Scholar ] [ CrossRef ]
Seyfried, T.N.; Sanderson, T.M.; El-Abbadi, M.M.; McGowan, R.; Mukherjee, P. Role of glucose and ketone bodies in the metabolic control of experimental brain cancer. Br. J. Cancer 2003 , 89 , 1375–1382. [ Google Scholar ] [ CrossRef ]
Maurer, G.D.; Brucker, D.P.; Bähr, O.; Harter, P.N.; Hattingen, E.; Walenta, S.; Mueller-Klieser, W.; Steinbach, J.P.; Rieger, J. Differential utilization of ketone bodies by neurons and glioma cell lines: A rationale for ketogenic diet as experimental glioma therapy. BMC Cancer 2011 , 11 , 315. [ Google Scholar ] [ CrossRef ]
Rieger, J.; Bähr, O.; Maurer, G.D.; Hattingen, E.; Franz, K.; Brucker, D.; Walenta, S.; Kämmerer, U.; Coy, J.F.; Weller, M.; et al. ERGO: A pilot study of ketogenic diet in recurrent glioblastoma. Int. J. Oncol. 2014 , 44 , 1843–1852. [ Google Scholar ] [ CrossRef ]
Abdelwahab, M.G.; Fenton, K.E.; Preul, M.C.; Rho, J.M.; Lynch, A.; Stafford, P.; Scheck, A.C. The ketogenic diet is an effective adjuvant to radiation therapy for the treatment of malignant glioma. PLoS ONE 2012 , 7 , e36197. [ Google Scholar ] [ CrossRef ]
Zhou, W.; Mukherjee, P.; Kiebish, M.A.; Markis, W.T.; Mantis, J.G.; Seyfried, T.N. The calorically restricted ketogenic diet, an effective alternative therapy for malignant brain cancer. Nutr. Metab. 2007 , 4 , 5. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Ciusani, E.; Vasco, C.; Rizzo, A.; Girgenti, V.; Padelli, F.; Pellegatta, S.; Fariselli, L.; Bruzzone, M.G.; Salmaggi, A. MR-Spectroscopy and Survival in Mice with High Grade Glioma Undergoing Unrestricted Ketogenic Diet. Nutr. Cancer 2021 , 73 , 2315–2322. [ Google Scholar ] [ CrossRef ]
De Feyter, H.M.; Behar, K.L.; Rao, J.U.; Madden-Hennessey, K.; Ip, K.L.; Hyder, F.; Drewes, L.R.; Geschwind, J.F.; De Graaf, R.A.; Rothman, D.L. A ketogenic diet increases transport and oxidation of ketone bodies in RG2 and 9L gliomas without affecting tumor growth. Neuro Oncol. 2016 , 18 , 1079–1087. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.; Brennan, S.E.; et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ 2021 , 372 , n71. [ Google Scholar ] [ CrossRef ]
Nebeling, L.C.; Miraldi, F.; Shurin, S.B.; Lerner, E. Effects of a ketogenic diet on tumor metabolism and nutritional status in pediatric oncology patients: Two case reports. J. Am. Coll. Nutr. 1995 , 14 , 202–208. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Zuccoli, G.; Marcello, N.; Pisanello, A.; Servadei, F.; Vaccaro, S.; Mukherjee, P.; Seyfried, T.N. Metabolic management of glioblastoma multiforme using standard therapy together with a restricted ketogenic diet: Case Report. Nutr. Metab. 2010 , 7 , 33. [ Google Scholar ] [ CrossRef ]
Han, L.; Zhang, J.; Zhang, P.; Han, X.; Bi, Z.; Li, J.; Wang, L.; Xue, F.; Fan, Y. Perspective research of the influence of caloric restriction combined with psychotherapy and chemotherapy associated by hybaroxia on the prognosis of patients suffered by glioblastoma multiforme. Zhonghua Yi Xue Za Zhi 2014 , 94 , 2129–2131. [ Google Scholar ] [ PubMed ]
Schwartz, K.; Chang, H.T.; Nikolai, M.; Pernicone, J.; Rhee, S.; Olson, K.; Kurniali, P.C.; Hord, N.G.; Noel, M. Treatment of glioma patients with ketogenic diets: Report of two cases treated with an IRB-approved energy-restricted ketogenic diet protocol and review of the literature. Cancer Metab. 2015 , 3 , 3. [ Google Scholar ] [ CrossRef ]
Champ, C.E.; Palmer, J.D.; Volek, J.S.; Werner-Wasik, M.; Andrews, D.W.; Evans, J.J.; Glass, J.; Kim, L.; Shi, W. Targeting metabolism with a ketogenic diet during the treatment of glioblastoma multiforme. J. Neurooncol. 2014 , 117 , 125–131. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Santos, J.G.; Da Cruz, W.M.; Schönthal, A.H.; Salazar, M.D.; Fontes, C.A.; Quirico-Santos, T.; Da Fonseca, C.O. Efficacy of a ketogenic diet with concomitant intranasal perillyl alcohol as a novel strategy for the therapy of recurrent glioblastoma. Oncol. Lett. 2018 , 15 , 1263–1270. [ Google Scholar ] [ CrossRef ]
Van der Louw, E.; Reddingius, R.E.; Olieman, J.F.; Neuteboom, R.F.; Catsman-Berrevoets, C.E. Ketogenic diet treatment in recurrent diffuse intrinsic pontine glioma in children: A safety and feasibility study. Pediatr. Blood Cancer 2019 , 66 , e27561. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Martin-McGill, K.J.; Marson, A.G.; Tudur Smith, C.; Jenkinson, M.D. The Modified Ketogenic Diet in Adults with Glioblastoma: An Evaluation of Feasibility and Deliverability within the National Health Service. Nutr. Cancer 2018 , 70 , 643–649. [ Google Scholar ] [ CrossRef ] [ PubMed ]
van der Louw, E.J.; Olieman, J.F.; van den Bemt, P.M.; Bromberg, J.E.; Oomen-de Hoop, E.; Neuteboom, R.F.; Catsman-Berrevoets, C.E.; Vincent, A.J. Ketogenic diet treatment as adjuvant to standard treatment of glioblastoma multiforme: A feasibility and safety study. Ther. Adv. Med. Oncol. 2019 , 11 , 1758835919853958. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Woodhouse, C.; Ward, T.; Gaskill-Shipley, M.; Chaudhary, R. Feasibility of a modified Atkins diet in glioma patients during radiation and its effect on radiation sensitization. Curr. Oncol. 2019 , 26 , e433–e438. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Martin-McGill, K.J.; Marson, A.G.; Tudur Smith, C.; Young, B.; Mills, S.J.; Cherry, M.G.; Jenkinson, M.D. Ketogenic diets as an adjuvant therapy for glioblastoma (KEATING): A randomized, mixed methods, feasibility study. J. Neurooncol. 2020 , 147 , 213–227. [ Google Scholar ] [ CrossRef ]
Klein, P.; Tyrlikova, I.; Zuccoli, G.; Tyrlik, A.; Maroon, J.C. Treatment of glioblastoma multiforme with “classic” 4:1 ketogenic diet total meal replacement. Cancer Metab. 2020 , 8 , 24. [ Google Scholar ] [ CrossRef ]
Panhans, C.M.; Gresham, G.; Amaral, L.J.; Hu, J. Exploring the Feasibility and Effects of a Ketogenic Diet in Patients With CNS Malignancies: A Retrospective Case Series. Front. Neurosci. 2020 , 14 , 390. [ Google Scholar ]
Voss, M.; Wagner, M.; von Mettenheim, N.; Harter, P.N.; Wenger, K.J.; Franz, K.; Bojunga, J.; Vetter, M.; Gerlach, R.; Glatzel, M.; et al. ERGO2: A Prospective, Randomized Trial of Calorie-Restricted Ketogenic Diet and Fasting in Addition to Reirradiation for Malignant Glioma. Int. J. Radiat. Oncol. Biol. Phys. 2020 , 108 , 987–995. [ Google Scholar ] [ CrossRef ]
Schreck, K.C.; Hsu, F.C.; Berrington, A.; Henry-Barron, B.; Vizthum, D.; Blair, L.; Kossoff, E.H.; Easter, L.; Whitlow, C.T.; Barker, P.B.; et al. Feasibility and Biological Activity of a Ketogenic/Intermittent-Fasting Diet in Patients With Glioma. Neurology 2021 , 97 , e953–e963. [ Google Scholar ] [ CrossRef ]
Perez, A.; van der Louw, E.; Nathan, J.; El-Ayadi, M.; Golay, H.; Korff, C.; Ansari, M.; Catsman-Berrevoets, C.; von Bueren, A.O. Ketogenic diet treatment in diffuse intrinsic pontine glioma in children: Retrospective analysis of feasibility, safety, and survival data. Cancer Rep. 2021 , 4 , e1383. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Seyfried, T.N.; Shivane, A.G.; Kalamian, M.; Maroon, J.C.; Mukherjee, P.; Zuccoli, G. Ketogenic Metabolic Therapy, Without Chemo or Radiation, for the Long-Term Management of IDH1-Mutant Glioblastoma: An 80-Month Follow-Up Case Report. Front. Nutr. 2021 , 8 , 682243. [ Google Scholar ] [ CrossRef ]
Porper, K.; Shpatz, Y.; Plotkin, L.; Pechthold, R.G.; Talianski, A.; Champ, C.E.; Furman, O.; Shimoni-Sebag, A.; Symon, Z.; Amit, U.; et al. A Phase I clinical trial of dose-escalated metabolic therapy combined with concomitant radiation therapy in high-grade glioma. J. Neurooncol. 2021 , 153 , 487–496. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Voss, M.; Wenger, K.J.; von Mettenheim, N.; Bojunga, J.; Vetter, M.; Diehl, B.; Franz, K.; Gerlach, R.; Ronellenfitsch, M.W.; Harter, P.N.; et al. Short-term fasting in glioma patients: Analysis of diet diaries and metabolic parameters of the ERGO2 trial. Eur. J. Nutr. 2022 , 61 , 477–487. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Phillips, M.C.; Leyden, J.; McManus, E.J.; Lowyim, D.G.; Ziad, F.; Moon, B.G.; Haji Mohd Yasin, N.A.; Tan, A.; Thotathil, Z.; Jameson, M.B. Feasibility and Safety of a Combined Metabolic Strategy in Glioblastoma Multiforme: A Prospective Case Series. J. Oncol. 2022 , 2022 , 4496734. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Schwartz, K.A.; Noel, M.; Nikolai, M.; Olson, L.K.; Hord, N.G.; Zakem, M.; Clark, J.; Elnabtity, M.; Figueroa, B.; Chang, H.T. Long Term Survivals in Aggressive Primary Brain Malignancies Treated With an Adjuvant Ketogenic Diet. Front. Nutr. 2022 , 9 , 770796. [ Google Scholar ] [ CrossRef ]
Phillips, M.C.L.; Thotathil, Z.; Dass, P.H.; Ziad, F.; Moon, B.G. Ketogenic metabolic therapy in conjunction with standard treatment for glioblastoma: A case report. Oncol. Lett. 2024 , 27 , 230. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Schmidt, M.; Pfetzer, N.; Schwab, M.; Strauss, I.; Kammerer, U. Effects of a ketogenic diet on the quality of life in 16 patients with advanced cancer: A pilot trial. Nutr. Metab. 2011 , 8 , 54. [ Google Scholar ] [ CrossRef ]
Feasibility Study of Modified Atkins Ketogenic Diet in the Treatment of Newly Diagnosed Malignant Glioma. 2017. Available online: https://clinicaltrials.gov/study/NCT03278249 (accessed on 1 June 2024).
Ketogenic Diet Adjunctive to Salvage Chemotherapy for Recurrent Glioblastoma: A Pilot Study. 2016. Available online: https://clinicaltrials.gov/study/NCT02939378 (accessed on 1 June 2024).
A Pilot Study of Ketogenic Diet and Metformin in Glioblastoma: Feasibility and Metabolic Imaging. 2020. Available online: https://clinicaltrials.gov/study/NCT04691960 (accessed on 1 June 2024).
Pilot Study of a Metabolic Nutritional Therapy for the Management of Primary Brain Tumors. 2012. Available online: https://clinicaltrials.gov/study/NCT01535911 (accessed on 1 June 2024).
A Randomized Controlled Phase 2 Study of the Ketogenic Diet Versus Standard Dietary Guidance for Patients With Newly Diagnosed Glioblastoma in Combination with Standard-Of-Care Treatment. 2023. Available online: https://clinicaltrials.gov/study/NCT05708352 (accessed on 1 June 2024).
A Phase 2 Trial of Paxalisib Combined with a Ketogenic Diet and Metformin for Newly Diagnosed and Recurrent Glioblastoma. 2021. Available online: https://clinicaltrials.gov/study/NCT05183204 (accessed on 1 June 2024).
A “Classic” Ketogenic Diet as a Complementary Therapeutic Management on Patients with High-Grade Gliomas and Brain Metastases. 2022. Available online: https://clinicaltrials.gov/study/NCT05564949 (accessed on 1 June 2024).
IIT2016-17-HU-KETORADTMZ: A Phase 1 Study of a 4-Month Ketogenic Diet in Combination with Standard-Of-Care Radiation and Temozolomide for Patients With Newly/Recently Diagnosed Glioblastoma. 2018. Available online: https://clinicaltrials.gov/study/NCT03451799 (accessed on 1 June 2024).
Feasibility, Safety, and Efficacy of a Metabolic Therapy Program in Conjunction with Standard Treatment for Glioblastoma. 2020. Available online: https://clinicaltrials.gov/study/NCT04730869 (accessed on 1 June 2024).
Clontz, A.D. Ketogenic therapies for glioblastoma: Understanding the limitations in transitioning from mice to patients. Front. Nutr. 2023 , 10 , 1110291. [ Google Scholar ] [ CrossRef ]
Ebrahimpour-Koujan, S.; Shayanfar, M.; Benisi-Kohansal, S.; Mohammad-Shirazi, M.; Sharifi, G.; Esmaillzadeh, A. Adherence to low carbohydrate diet in relation to glioma: A case-control study. Clin. Nutr. 2019 , 38 , 2690–2695. [ Google Scholar ] [ CrossRef ]
Noorlag, L.; De Vos, F.Y.; Kok, A.; Broekman, M.L.; Seute, T.; Robe, P.A.; Snijders, T.J. Treatment of malignant gliomas with ketogenic or caloric restricted diets: A systematic review of preclinical and early clinical studies. Clin. Nutr. 2019 , 38 , 1986–1994. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Marinescu, S.C.; Apetroaei, M.M.; Nedea, M.I.; Arsene, A.L.; Velescu, B.Ș.; Hîncu, S.; Stancu, E.; Pop, A.L.; Drăgănescu, D.; Udeanu, D.I. Dietary Influence on Drug Efficacy: A Comprehensive Review of Ketogenic Diet-Pharmacotherapy Interactions. Nutrients 2024 , 16 , 1213. [ Google Scholar ] [ CrossRef ]
Wang, Y.; Jing, M.X.; Jiang, L.; Jia, Y.F.; Ying, E.; Cao, H.; Guo, X.Y.; Sun, T. Does a ketogenic diet as an adjuvant therapy for drug treatment enhance chemotherapy sensitivity and reduce target lesions in patients with locally recurrent or metastatic Her-2-negative breast cancer? Study protocol for a randomized controlled trial. Trials 2020 , 21 , 487. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Iyikesici, M.S. Survival outcomes of metabolically supported chemotherapy combined with ketogenic diet, hyperthermia, and hyperbaric oxygen therapy in advanced gastric cancer. Niger. J. Clin. Pract. 2020 , 23 , 734–740. [ Google Scholar ] [ CrossRef ]
Yang, L.; TeSlaa, T.; Ng, S.; Nofal, M.; Wang, L.; Lan, T.; Zeng, X.; Cowan, A.; McBride, M.; Lu, W.; et al. Ketogenic diet and chemotherapy combine to disrupt pancreatic cancer metabolism and growth. Med 2022 , 3 , 119–136. [ Google Scholar ] [ CrossRef ]
Klement, R.J.; Schafer, G.; Sweeney, R.A. A ketogenic diet exerts beneficial effects on body composition of cancer patients during radiotherapy: An interim analysis of the KETOCOMP study. J. Tradit. Complement. Med. 2020 , 10 , 180–187. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Caffa, I.; Spagnolo, V.; Vernieri, C.; Valdemarin, F.; Becherini, P.; Wei, M.; Brandhorst, S.; Zucal, C.; Driehuis, E.; Ferrando, L.; et al. Fasting-mimicking diet and hormone therapy induce breast cancer regression. Nature 2020 , 583 , 620–624. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Meidenbauer, J.J.; Mukherjee, P.; Seyfried, T.N. The glucose ketone index calculator: A simple tool to monitor therapeutic efficacy for metabolic management of brain cancer. Nutr. Metab. 2015 , 12 , 12. [ Google Scholar ] [ CrossRef ]
De Groot, S.; Pijl, H.; van der Hoeven, J.J.M.; Kroep, J.R. Effects of short-term fasting on cancer treatment. J. Exp. Clin. Cancer Res. 2019 , 38 , 209. [ Google Scholar ] [ CrossRef ]
Klement, R.J. Fasting, Fats, and Physics: Combining Ketogenic and Radiation Therapy against Cancer. Complement. Med. Res. 2018 , 25 , 102–113. [ Google Scholar ] [ CrossRef ]
Demirel, A.; Li, J.; Morrow, C.; Barnes, S.; Jansen, J.; Gower, B.; Kirksey, K.; Redden, D.; Yarar-Fisher, C. Evaluation of a ketogenic diet for improvement of neurological recovery in individuals with acute spinal cord injury: Study protocol for a randomized controlled trial. Trials 2020 , 21 , 372. [ Google Scholar ] [ CrossRef ]
Bough, K.J.; Yao, S.G.; Eagles, D.A. Higher ketogenic diet ratios confer protection from seizures without neurotoxicity. Epilepsy Res. 2000 , 38 , 15–25. [ Google Scholar ] [ CrossRef ]
Titcomb, T.J.; Liu, B.; Wahls, T.L.; Snetselaar, L.G.; Shadyab, A.H.; Tabung, F.K.; Saquib, N.; Arcan, C.; Tinker, L.F.; Wallace, R.B.; et al. Comparison of the Ketogenic Ratio of Macronutrients With the Low-Carbohydrate Diet Score and Their Association With Risk of Type 2 Diabetes in Postmenopausal Women: A Secondary Analysis of the Women’s Health Initiative. J. Acad. Nutr. Diet. 2023 , 123 , 1152–1161.e4. [ Google Scholar ] [ CrossRef ]
Wirrell, E.C. Ketogenic ratio, calories, and fluids: Do they matter? Epilepsia 2008 , 49 (Suppl. S8), 17–19. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Zilberter, T.; Zilberter, Y. Ketogenic Ratio Determines Metabolic Effects of Macronutrients and Prevents Interpretive Bias. Front. Nutr. 2018 , 5 , 75. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Li, S.; Du, Y.; Meireles, C.; Sharma, K.; Qi, L.; Castillo, A.; Wang, J. Adherence to ketogenic diet in lifestyle interventions in adults with overweight or obesity and type 2 diabetes: A scoping review. Nutr. Diabetes 2023 , 13 , 16. [ Google Scholar ] [ CrossRef ]
Shim, J.S.; Oh, K.; Kim, H.C. Dietary assessment methods in epidemiologic studies. Epidemiol. Health 2014 , 36 , e2014009. [ Google Scholar ] [ CrossRef ]
Anghel, L.A.; Farcas, A.M.; Oprean, R.N. An overview of the common methods used to measure treatment adherence. Med. Pharm. Rep. 2019 , 92 , 117–122. [ Google Scholar ] [ CrossRef ] [ PubMed ]
Iwendi, C.; Khan, S.; Anajemba, J.H.; Bashir, A.K.; Noor, F.J.I.A. Realizing an Efficient IoMT-Assisted Patient Diet Recommendation System Through Machine Learning Model. IEEE Access 2020 , 8 , 28462–28474. [ Google Scholar ] [ CrossRef ]

Click here to enlarge figure

Study	Type of Study	Number of Participants	Overall Survival
Nebeling et al., 1995 [ ]	Case Report	2	60 months and 48 months
Zuccoli et al., 2010 [ ]	Case Report	1	NA
Han et al., 2014 [ ]	Prospective Study	11	Mean Survival: 38 +/− 13 months
Schwartz et al., 2015 [ ]	Case Report	2	NA
Rieger et al., 2014 [ ]	Prospective Study	20	Median: 32 weeks
Champ et al., 2014 [ ]	Retrospective Analysis	134	Median: 14 months
Santos et al., 2017 [ ]	Prospective Randomized Study	37	NA
Van der Louw et al., 2018 [ ]	Prospective Study	3	16.5, 6.4 and 18.7 months
Martin-McGill et al., 2018 [ ]	Prospective Study	6	NA
Van der Louw et al., 2019 [ ]	Prospective Study	11	9.8 and 19.0 months
Woodhouse et al., 2019 [ ]	Retrospective Study	29	Not evaluated
Martin-McGill et al., 2020 [ ]	Prospective Study	12	Median: 67.3 weeks
Klein et al., 2020 [ ]	Prospective Randomized Study	8	Group 1: 20 months (9.5–27) Group 2: 12.8 months (6.3–19.9)
Panhans et al., 2020 [ ]	Retrospective Case Series	12	90.8–19.0 Months
Voss et al., 2020 [ ]	Prospective Randomized Study	50	KD: 331 days SD: 291 days Low glucose KD: 348 days
Schreck et al., 2021 [ ]	Prospective Study	25	NA
Perez et al., 2021 [ ]	Retrospective Study	5	Median: 18.7 Months
Seyfried et al., 2021 [ ]	Case Report	1	80 Months
Porper et al., 2021 [ ]	Prospective Randomized Study	13	21 Months in patients with newly diagnosed disease 8 Months in patients with recurrent disease
Voss et al., 2022 [ ]	Prospective Randomized Study	50	250–485 days
Phillips et al., 2022 [ ]	Prospective Case Series	10	Median: 13 Months
Schwartz et al., 2022 [ ]	Prostective Study	12	Not Reported
Philli s et al., 2024 [ ]	Case Report	1	36 Months

Study	Adverse Events Related to KD
Nebelin et al., 1995 [ ]	No reported symptoms
Zuccoli et al., 2010 [ ]	hyperuricemia
Han et al., 2014 [ ]	N/A
Schwartz et al., 2015 [ ]	No significant adverse events
Rieger et al., 2014 [ ]	Weight loss, diarrhea, constipation, hunger
Champ et al., 2014 [ ]	Constipation, asthenia, weight loss, nephrolithiasis, hypoglycemia
Santos et al., 2017 [ ]	Not reported
Van der Louw et al., 2018 [ ]	Hypoglycemia, hyperkeratosis, vomiting, refusal to eat, asthenia, constipation
Martin-McGill et al., 2018 [ ]	Constipation
van der Louw et al., 2019 [ ]	Constipation, nausea/vomiting, hypercholesterolemia, hypoglycemia, diarrhea, low carnitine concentration
Woodhouse et al., 2019 [ ]	Grade 2 constipation occurred in 1 patient. Grade 1 fatigue and nausea probably due to standard therapy
Martin-McGill et al., 2020 [ ]	Hypokalemia, hypocalcemia, hypernatremia, hyperkalemia, constipation
Klein et al., 2020 [ ]	Weight loss, hunger, nausea, dizziness, asthenia, constipation
Panhans et al., 2020 [ ]	Asthenia, weight loss, nausea, vomiting, headache, decreased appetite
Voss et al., 2020 [ ]	Epileptic seizures, headache, nausea
Schreck et al., 2021 [ ]	Grade 2 adverse events: Leukopenia, nausea, diarrhea, fatigue, or seizure. Grade 3: neutropenia (possibly related)
Perez et al., 2021 [ ]	Hypoglycemia, constipation, hyperkeratosis, vomiting asthenia, hyperuricemia
Seyfried et al., 2021 [ ]	Not reported
Porper et al., 2021 [ ]	Nausea, asymptomatic hyperuricemia, anorexia
Voss et al., 2022 [ ]	Gastrointestinal symptoms, headache, muscle cramps
Phillips et al., 2022 [ ]	Fatigue, irritability, and feeling lightheaded. No grade 3 or higher adverse events.
Schwartz et al., 2022 [ ]	Not reported.
Phillips et al., 2024 [ ]	Prolonged fasts caused mild fatigue, diarrhea, and cold intolerance. No adverse events for KD.

The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

Valerio, J.; Borro, M.; Proietti, E.; Pisciotta, L.; Olarinde, I.O.; Fernandez Gomez, M.; Alvarez Pinzon, A.M. Systematic Review and Clinical Insights: The Role of the Ketogenic Diet in Managing Glioblastoma in Cancer Neuroscience. J. Pers. Med. 2024 , 14 , 929. https://doi.org/10.3390/jpm14090929

Valerio J, Borro M, Proietti E, Pisciotta L, Olarinde IO, Fernandez Gomez M, Alvarez Pinzon AM. Systematic Review and Clinical Insights: The Role of the Ketogenic Diet in Managing Glioblastoma in Cancer Neuroscience. Journal of Personalized Medicine . 2024; 14(9):929. https://doi.org/10.3390/jpm14090929

Valerio, Jose, Matteo Borro, Elisa Proietti, Livia Pisciotta, Immanuel O. Olarinde, Maria Fernandez Gomez, and Andres Mauricio Alvarez Pinzon. 2024. "Systematic Review and Clinical Insights: The Role of the Ketogenic Diet in Managing Glioblastoma in Cancer Neuroscience" Journal of Personalized Medicine 14, no. 9: 929. https://doi.org/10.3390/jpm14090929

Article Metrics

Article access statistics, further information, mdpi initiatives, follow mdpi.

Subscribe to receive issue release notifications and newsletters from MDPI journals

Share full article

On Children, Meaning, Media and Psychedelics

The writer jia tolentino on parenting — and living a good life — in the age of smartphones..

[MUSIC PLAYING]

From New York Times Opinion, this is “The Ezra Klein Show.”

And now for something completely different. We recorded this episode right before the first presidential debate, and there has been such a crush of political news since then, but there hasn’t really been a moment that felt right to release it. But I loved this conversation. And in a funny way, it’s more relevant now, given how much the election has come to revolve around the reasons people do and don’t have children and the meaning of that choice.

So a few months ago, Jia Tolentino published a big piece in The New Yorker” on “CoComelon.” “CoComelon,” if you do not have a two-year-old, is a show that every really little kid really loves and every parent has a more complicated set of emotions about. But it’s something Tolentino wrote at the end that was what really caught my eye. She said, “I found myself wondering if we’d be better off thinking less about educational value in children’s media and more about real pleasure, both for us and for our kids.”

In a way, this is an episode about real pleasure, which is not what I went into it thinking it would be about. It’s about the tension between pursuing pleasure, or what I might call meaning, and pursuing the kinds of achievements we spend most of our lives being taught to prize. Honestly, I think this gets much more to the heart of the questions people ask about having children than all this political rhetoric about cat ladies and extra votes and tax rates. And I don’t think it’s an accident that in this conversation, as we’re trying to talk about the value of what we can’t measure against the value of what we can, we end up finding ourselves in the language of religion, of psychedelics, of emotion. These are questions where I think we’ve culturally lost some of the vocabulary that we used to have to talk about just what it means to live a good life. Not to have a higher income or a better job, but what is a good life.

Jia Tolentino is the author of the great book of essays, “Trick Mirror,” one of my favorite books about being alive in the age of the internet. She’s a staff writer at The New Yorker. And as always, my email is [email protected]

Jia Tolentino, welcome to the show.

Thank you for having me back.

So, you told me that you came to a new understanding of why you had children on your way to tape today outside the Port Authority, and you would tell it to me when we were on the show. So, why did you have children?

I was thinking on the train up here about that question, like why did I have kids? And I was thinking about my trepidation beforehand. And I feel like I bring back every conversation about children to a conversation about psychedelics, unfortunately. But the idea seemed scary and overwhelming in the same way that doing acid seemed scary and overwhelming before I did it for the first time. It was like, oh, this is going to last for so long. There’s going to be part of it that’s so intense and so difficult. And I didn’t do it until I felt like I know that the person that I’m going to do it with, I’ll have fun with. I can trust that I’m doing it in kind of a safe and right environment, where I will get the thing that I want out of it.

But the thing that made me decide to do acid for the first time is not dissimilar to the thing that made me decide to have kids, which is, I think it’ll be fun. I think on the whole, I think it’ll be fun. I felt that there would be real, lasting, kind of destabilizing, kind of boundary-dissolving pleasure in it that would kind of scare me in the way that true pleasure kind of does.

And I really hadn’t thought about it that neatly until you said that we wanted to talk about this. I don’t think I understood that, really, the thing that drove me to this was probably the thing that drives me to a lot of things, which is pleasure-seeking.

It’s funny because I sometimes use the psychedelics and parenthood analogy, but I use it in a very different way, which is people will tell me they’re struggling with the decision, and they’re reading the parenting books. And I always say that to read the parenting books is, it’s like the difference between reading about doing psychedelics and doing psychedelics —

— and that the fun is not the point. I have this discomfort with the discourse around fun and parenting, as if the way to measure any experience in your life is whether, when you’re filling out a time use survey, you’re having a lot of fun doing it.

When I’ve done psychedelics, I don’t necessarily think they’re fun. Sometimes they are. But what brings me to them is meaning. And I feel like what brought me to parenting or what attracted me to it, what made it seem not even a question to me was that I want meaning in my life.

And I mean, what is a more fundamental sense of human meaning than continuing the human chain?

Well, I should say, too, when I say fun — I mean, I think that’s why I corrected myself. What I think of as fun is, it’s much less enjoyment and more like pushing the limits of what — I don’t know — can stand or I’m capable of. I have a kind of arduous idea of fun.

Like, something that I long to do constantly is go to Antarctica and completely lose my mind. That sounds like one of the most fun things I can imagine. And so, doing psychedelics, it is extremely challenging sometimes and not always fun, but that is a specific kind of pleasure that the definition of which is very close to finding meaning.

One thing that’s very clear to me, as your work has shifted towards thinking a lot about parenting, I think since you’ve become a parent, is that you find it really interesting. It’s very intellectually generative for you. I find that’s true for me. I think that the thing I always say to people about parenting that was surprising to me is how interesting it is. It was really undersold to me how just kind of it would focus my mind on things I would have never thought of or never thought of at that depth before.

You’ve written one of my favorite lines about parenting, and you wrote it in this piece about Angela Garbes’ book and about your own experience hiring a nanny. And you wrote, “We could afford to do this because a person can get paid more to sit in front of a computer and send a bunch of emails than she can to do a job that’s so crucial and difficult that it seems objectively holy: to clean excrement off a body, to hold a person while they are crying, to cherish them because of and not despite their vulnerability.” Tell me about the choice of the word “holy” there.

It’s the only — part of this is because the way I was raised, deep in the evangelical church in Texas, but that’s the only word for it, you know? We were talking about fun. We were talking about pleasure. Now we’re talking about this idea of the sacred, right? And I think that for me, the thing connecting all of those is some sort of submission and disappearance into something, right? It’s the total submission to someone else’s body, really, in your baby, you know? That, I found there’s no other word for it and to their bodies needs and to the mess of it and the — yeah, there’s no other word.

And it also feels the same way with the parts of parenting that are, in fact, tedious and repetitive and so mundane, which is often the exact same stuff, right? Like wiping a butt over and over again and wiping spit-up from somebody’s mouth and washing tiny, little hands, right? These things are so often tedious, and they are holy. And the thing that connects them both, it’s submission.

And I found that the transcendent moments in parenting and the really just objectively boring ones, where I’m laying on the floor of my living room wishing I could read a book, instead of just stacking little plastic eggs on each other, it feels like the same project.

And I have found parenting really interesting. And I think by the time I decided to try to do it, I figured it would be, right? I figured it would be in the same way that I was like, no matter what, as with my first acid trip, it was like, no matter what, this will be interesting. No matter what, this will be extremely difficult in a way that is interesting. How could it not?

I think there’s so much tension and energy and guilt in this connection between the sacred and the mundane, the sense that you often should be feeling. You are so close to this transcendent experience. You are doing the most meaningful thing, and you are so bored or so tired, or you so want to be somewhere else.

What do you do when what you are trying to do is escape the thing you should be paying attention to? And it’s such a profound, constant experience in parenting but also in life.

Right? To be alive is also holy. To be alive, to be able to experience this at any moment, right? The possibility of connection, of experience, of just being in this world, it should be so overwhelming. And instead, I am staring at my phone.

Phone, yeah. I started laughing when you were talking about that because I was just thinking about my little baby’s about to turn one. And so it’s like, I just instantly thought back to — I don’t know — days when she was like four months old. And you know those days when it’s like 9:30 a.m., and you’re like, I wish everyone would go to bed, you know? Anyone ever have those days?

And I would feel, when I had that thought, right, like being tired from getting up in the middle of the night, whatever it was, and I was just like, can everyone just go to bed now so I can not speak and not do anything and not learn and not play or whatever?

And I would have the thought, I’m abrogating the whole purpose of being alive, right? I could actually just enjoy this, pushing the swing in the sunshine over and over and over. But instead, I just want to look at some dumb shit on my phone and whatever.

And — I don’t know — have you ever had this experience? I think back to when I was little, and I would read books while I was on roller skates. I don’t think that the quality of wanting to leap out of the texture of the present is something that’s specific to the smartphone era.

I remember I read in the bathtub. I just always — you memorized the shampoo bottles. You’re always kind of looking for a narrative to take you out of the present. And I remember that being something that was true for me as a really little kid, even as I was someone and remain someone that — I mean, I was very present, and I did have a great time all the time, pretty much.

But I do think that the way that the smartphone has sort of deformed and put that desire quivering in our pockets, beckoning to us, I was going to say, I don’t know if this has ever happened to you, but I have this sort of symptom of this brain disease that’s particularly troubling to me, is like, pre-kids, back when I had enough alone time to have original thoughts more than once every five months or something, I would think something.

And I would have this sense of, this is an idea that is shimmering with movement, in some way, you know? And then it would be too much for me. And then I would be like, I can’t. I can’t deal with it. I’m going to write it down, and then I’m going to scroll for five minutes. Like I would very frequently have that response. And that terrified me even though I kept having it.

And it sometimes feels to me not that we’re turning away from the mess and the wonder of real physical experience, despite the fact that it’s precious. I kind of feel something within me sometimes that it’s too precious. It’s too much, that being present is work, in a way, that it’s this rawness, and it’s this mutability. It requires this of us and a presence. That is something that I have sometimes found myself flexing away from because of all the reasons that it’s good, in a weird way. Have you ever — do you know what I mean at all?

I absolutely know what you mean in a million different ways. I mean, I was a kid. Why do I read? I mean, now I think it’s almost a leftover habit, but I read to escape. I read to escape my world. I read to escape my family. I read to escape things I didn’t understand. And I read obsessively, constantly, all the time, in cars, in the bathroom, anywhere.

Because it was a socially sanctioned way to be alone.

And nobody would bother me because it was virtuous for me to be reading.

One of the things I was thinking about what you were saying that was we have — there’s more spiritualism in this conversation than I expected, but I’m enjoying it. And it feels a little bit like our metaphors are shaped by different traditions, right? I know you grew up deeply Christian. And there’s a sense in the way you think about and write about it — the holiness, the awe of all creation, right? The external world that requires something of you.

And a lot of my experience of this or thinking about it is shaped more from meditation and mindfulness. And so the thing that I was thinking about what you were saying that was what always feels limited to me is my attention. And a lot of the need to escape is a need to rest my attention and recharge it.

And what allows me to access the transcendence of my children, of the world is, honestly, how rested and how awake and aware I am. I mean, I spent some time in a coffee shop before coming here to talk to you, and I just needed that time listening to music and reading, so my attention could recover, so I could be present here with you.

So in that way, I think escape is undertheorized. That escape, it can be good or bad. I think we have trouble with this question of, are we distracting ourselves, or are we recovering?

Are we getting a kind of necessary contemplation, so that we can come back and experience a world and process what we’ve experienced and seen, or are we running from it? Are we trying not to feel from it? Are we trying to be anywhere but here?

Or are we looking for something, right? Are we looking — because I think that a lot of what takes people to screens is, it looks like escape, but I think it’s also pursuit.

A friend of mine was watching my baby when it was one of those school’s off, whatever. Andrew’s out of town, whatever.

And I was saying, OK, before you put her down for a nap, rock her in the rocking chair. Give her a couple of minutes. And then once you see the eyes blink really heavy, just dump her. And they were like, oh, that’s her phone scrolling time. And I was like, yeah, that’s her phone scrolling time, you know? Her brain just needs to — sometimes we’re just putting our brain on the static signal.

I think this is a place where there’s so much self-judgment, right? Are you escaping? Are you recovering? And then we put that judgment also on our children. And this gets us to “CoComelon.” Why don’t you describe, for someone who has never seen it and has no idea what that word means, what is “CoComelon“?

So “CoComelon” is one of the most successful entertainment franchises of all time, not just for children. And yet it’s something that if you haven’t changed a diaper in the last four years, you probably have no idea that it exists, you know?

(SINGING) Beans, beans

It’s time to eat your beans

(SINGING) Yes, yes, yes

I want to eat the beans

Beans are good for you

Yay, yay, yay

I love them, ooh

How to put this? It’s this. The backdrop of “CoComelon” is that major children’s animation companies did not make entertainment for babies and young toddlers because this was seen as sort of unethical. And kids can’t really learn from a screen at that age, so we’re not going to do it. And then YouTube was invented, and then the iPad was invented.

And suddenly iPad parenting, of which I certainly take part in, was instantiated as the way that suddenly we were all living.

I think half of 2 to 5-year-olds have their own mobile devices, which, again, my 4-year-old is one of them. And then, so, just, this land was wide open, this pristine farmland of just millions and millions and millions and millions of children whose attention could then be captured and monetized. And then you get all these things. Like, people know Ms. Rachel, probably.

Hi. Hello Can you say “mama“? Mama. Mama. Let’s sing it.

(SINGING) Mama

Let’s clap it. Mama. Let’s sign it. Mama. Good job.

But then there were all of these nursery rhyme channels where you would get just sing-song nursery rhymes and kind of squeaky-looking, mesmerizing, uncanny animation of just giant, bobble-headed, eternally smiling babies, and perfect, little worlds where the sun is always shining. And the parents are always around. And it’s rainbow popsicles, and it’s fort building, and just smiles and smiles and smiles and smiles, and the same words repeated over and over and over and over, and bright, clanging noises, and these things that are torturous for adults, but basically heavenly to whatever is going on in our babies’ brains.

(SINGING) Wash my hair

Doo, doo, doo, doo, doo, doo

Wash my hair

Wash my face

“CoComelon” is like — it’s so popular that as of some time ago, the daily viewers, where it was like 80 million daily viewers, which is as many people has watched the 2016 presidential debate between Trump and Clinton. And that’s just everyday viewers. And I think a conservative estimate would be that it’s watched for 200 billion minutes a year. And it’s just all the more remarkable because most of those viewers are basically pre-verbal. I’ve learned much more about “CoComelon” than I ever thought I was going to.

There’s something interesting in “CoComelon,” and I assume this maybe partially motivated your inquiry into it. There are two of them, I think — really “CoComelon” and “Blippi,” but we’ll focus on “CoComelon”— that exist at this absolute tension point of children adore it and adults hate it. There are other things adults don’t mind, right? “Sesame Street,” actually, most adults like. They remember it. We remember it. I enjoy it. You can’t really get that many kids to watch “Mister Rogers” now, but if you can. But parents like “Daniel Tiger.”

“Daniel Tiger.”

There’s all kinds of stuff that parents will love. “Bluey,” I think, in general, parents like “Bluey” more than their kids like “Bluey,”

But “CoComelon” is this one where, from a very young age your children go — and I feel like you could describe this two ways — completely vacant or completely focused in front of it. It’s either an experience of being totally filled or totally empty, and I can never quite tell which. And parents just — it drives them mad. Why this one? Why do you feel like there’s this unfathomable divergence between what the kids want here and what the parents want here?

Well, I think, as far as I can make sense of it, this is the first sort of — like, “CoComelon” ushered in a paradigm where children’s entertainment is not configured as entertainment, but as just raw, attentional capture. And I think that that’s why. It’s the people that work on it or that worked on it, they’ve been laying people off like crazy, despite these viewing numbers and the $3 billion parent company valuation.

But I do think that the people that created it are interested in providing pleasure and entertainment for the people that watch it. But the project of this company is attentional capture. And obviously, there’s significant overlap between attentional capture and entertainment. But I do think that we can — and now I kind think we should meaningfully differentiate them, maybe especially for kids. And I think that’s why.

I think you can feel it in the sort of bones of the stuff and the reaction parents have to it, you know? Even “Teletubbies,” college students love coming down from drugs and watching “Teletubbies.” Even silly baby entertainment, it can provide delight, which is not saying any of this has ever been pure, right? Children’s television has basically always existed as like an eternal toy commercial.

But even “GI Joe,” “My Little Pony,” basically, way back, “Mickey Mouse Club,” whatever, all of these things, they were configured as entertainment first. And there’s something about this that doesn’t feel like it’s configured as entertainment first. It feels like it’s just eyeball capture, just mining attention with a pickax into the parents’ eyes.

How much do you think the parental anger at “CoComelon”— and I will very much include myself here — is a kind of self-loathing, though, right?

100 percent.

I love the line you said — not entertainment, but raw attentional capture. And two things really jump out at me from that. One is that when you’re putting an 18-month or two-year-old in front of “CoComelon,” you’re usually doing it for a reason. You desperately have to get something done around the house. Their older brother is sick. You’re on a plane, right? There’s a reason you’re doing it.

And what you’re doing is trying to create raw attentional capture. If it did not completely capture them, it would not be serving the instrumental purpose you are using it for. I mean, they’re too young for entertainment, really, at least in the way we think about it in culture. So, one, it’s like we have asked this thing to provide a service, and we are mad at how well it provides it.

And then, two, there is this creepy analogy to ourselves. I mean, how much that we absorb and consume is not entertainment, but raw attentional capture? How well does that describe parts of Instagram or TikTok or even television that we binge, knowing that it has almost no nourishment to it?

But we just don’t want to think. We are asking it to provide an instrumental service, which is make the time go faster and make me disembodied because I don’t want to be here right now, having my holy life. I want to be completely absorbed in something else, something outside of myself.

Yeah, so one thing about “CoComelon,” right, like you said, they’re not doing anything new. It’s just that the audience is new, right? I talked to former writers who were telling me that they got a spreadsheet of all of the search words that toddlers were making their parents type in on YouTube Kids or whatever, and they would write episodes to those search terms, right? It was extremely S.E.O. targeted operation. And everything that I see has been algorithmically tailored to exactly what I want to be looking at as well, you know?

And I think I will say, I don’t have that much screen time anxiety about my kids, right? I’m like, y’all have plenty of resources, you know? Like, you were creative class children in Brooklyn. You are luckier than 99 percent of the global population. You’re going to be fine. I don’t care about your specific language. I don’t know. I don’t worry about screen time in a very specific way.

But I have a preemptive sorrow about the way that any ill can be instantly evaporated by putting my phone in front of my toddler and letting her — or not my toddler anymore — my four-year-old and letting her text emojis to my partner all day, you know, another one of my distraction tactics.

I get the sense that you’re going to be looking at screens so much. You’re going to be doing everything that I’m doing, but probably by hours and hours worse. And you, like me, are going to be unable to be just present without reaching for your phone after a certain number of minutes probably, right?

And your conception of what is possible is going to be limited to what is presented to you on that screen and your conceptions of what you want. And I feel like all of that screen time anxiety that I feel about her comes from my own sense that screens have already foreclosed a lot of that negative capability in my own life.

I want to expand on a line you gestured at there when you talked about what are we afraid of. And this is one of the places your piece really connected for me. I have this feeling, as I said earlier, that we’re under-specified on what we want and what we don’t want.

And you write, “When it comes to the shows we allow our children to watch, we are afraid of what exactly? That our kids’ capacity for deep thought will be blunted by compulsive screen use? That they will lose their ability to sit with the plain fact of existence, to pay attention to the world as it is, to conceive of new possibilities? That they’ll grow up to be just like us, only worse?”

And those all feel like things we’re afraid of and also maybe that they will never know any different. And I wonder about this, right? I mean, my kids, they will never remember a time before YouTube kids, right? They didn’t exist in a time before YouTube kids.

There has always been escape. There’s always been distraction. And the fact that it was not that good, I think was important. And I have trouble describing this. And I have trouble then making the distinctions based on it. But it’s like, I want my children to be able to escape the difficulty of reality. I think it’s important. I do it, too.

But somehow, I know when I do it in certain ways, it’s bad for me. And when I do it in other ways, it’s good for me. I don’t know why it is bad for me to look at my phone and good for me to read a magazine. But it is, and I can’t put it on a chart for you.

Can you not? It’s because one is surveilled, and the other isn’t, right?

I don’t think I care about the surveillance.

You don’t? But don’t you think that’s why one feels better than the other? You feel freer doing one than the other? You don’t feel your choices being sort of actively manipulated and shaped and constrained by an extremely bald profit structure? That, to me, feels — like you said, the escape was worse, but it also was an escape, right?

We were in our books in the back of the car, and nobody knew what was happening and what we were reading, except for us. Our parents would never know. There was no machine record of it whatsoever — even if we were writing, right, if we were doing the equivalent of what is, I think, widely and rightfully configured as unhealthy, like the 11-year-old girl on Instagram.

The way I was processing my life in narrative or whatever, or the way I was writing my life into its existence, was in a notebook, where no one could see it, and no one would ever profit from escalating or distorting it or testing it against anything. And so much of that seems tied, for me, to the lack of silent, invisible, constant surveillance.

So — and I mean this completely sincerely — I love how much that doesn’t resonate for me at all.

Oh, really?

Because it means that I’m having such a different experience of this. And so I want to have them both here. Because, one, I’m so close to a phone vegan. I’m so unbelievably annoying about what I have on my phone. So virtually nothing on my phone even can surveil me at this point.

Do you not have a browser?

I have a browser, but I don’t use it that much, except literally to look at Pitchfork music reviews, which is one of my favorite time wasting activities. But I have a lot of — I have The New Yorker app. I’m unbelievably annoying as a person.

But for me, the thing that I notice about a magazine, which I think is my favorite form of media, like full-stop, is, there is room for me to get interested and absorbed and let my attention move away from where I am at that moment. But it is not so absorbing or so grasping that my attention can’t shift back, that I’ve fully lost track of my body or my surroundings.

And I think this is one of the things that I want to get at. I am unnerved by how much we feel the need to net everything out, whether culture is good or bad, to these very measurable outcomes about school achievement or income in 20 years or teen mental health. And it feels to me like we’ve just lost the ability to make judgments based on sort of virtues and values about when things are good and bad, like whether it is better to read a book or look at TikTok, irrespective of —

— whether that shows up into —

Studies of educational achievement, yeah.

Exactly. I feel like we have lost self-confidence in making cultural judgments for ourselves, for society. as parents, we are so achievement oriented. I feel like you see this in the debate about Jon Haidt’s book, “The Anxious Generation,” that if we cannot show something on a chart, it’s like we cannot have the self-confidence to make a judgment about it.

Well, I think I was going to pitch you another possible reason why the magazine feels different from the phone, which is that, I mean, the phone, it is always entwined with usefulness, and your work email lingers. I mean, you are reachable, to be useful to someone when you were looking at your phone, even if you’re just reading the same book on your Kindle that you would be reading off of it, right?

I think real pleasure, you’re not — there’s nothing quantifiably achieved in it. You are not just un-surveilled, but you are not being useful to a goddamn person, you know, except the person directly in front of you, if there is one, right, or the people, the many people. And that has to do with what you’re talking about. The sort of Emily Oster-esque — not knocking her — I subscribe to her Substack — like, the reduction of everything to, what does the data say, and what are the outcomes? And that’s the choice, right? It’s sort of maybe an intellectual inadequacy in my own life that causes me to come back to this.

But it’s like, does it feel good, or does it not? And I think we still know what feels good and bad. And I think that that’s a good of a metric to judge anything on for ourselves and our kids, to some extent, right?

I think that really helps me describe or realize something. I want to go back to what we were talking about a few minutes ago, this question of, what are you afraid of, afraid of for your children, specifically? And I think maybe that gets at it.

The reason I like magazines and don’t like my phone is not surveillance or anything else. It’s that I feel better when I read a magazine. There are a million reasons for that. And I would probably describe it as my attention is more collected and centered and stable afterwards, and so I can then attend to other things in my life better and with more joy. But it’s just I feel better. I like it. It’s more pleasurable.

And I think the thing that I worry about, the thing that I’m afraid of, is, I wanted to bring children into the world not because the world is perfect. But it is beautiful. And I feel like I got this gift of getting to experience it. And I want them to have this gift of getting to experience it in its wonderful dimensions and its horrors.

And I am worried that I have unleashed a set of technologies upon them — and that we’ve done this socially — that is going to structurally and permanently degrade their capacity for that experience, in the way that I notice it degrading mine in the moment. And I’m frustrated at myself for not being better at policing. But them, they’re young, and they’re getting tuned and trained and wired or whatever metaphor you want here.

And I think the thing I am afraid of is not like their grades will be bad because they watched “Ninjago.” It is that their experience of the world will be thinner and more scattered because they will have been trained on these hyper stimulating things that somehow absorb you at the same time they make you feel bad. And that will just become the sort of baseline of what attention is supposed to be like.

So they’ll be like us, right? They will have their experience of the world curtailed by the desire to check a device every however many seconds, right? They will. There’s just no doubt about it. It’ll probably be a lot worse for them. But hopefully, like us, too, they will have found the things in the world that they can be devoted to in a way that supersedes, at least a certain amount of time, the screen, right?

The only way that I see out of this for my children is the only way that I see out of this for myself, is like, I can’t be disciplined. I can’t spend X amount of minutes less per day on my phone because I know it’s bad for me how much I’m on there already. I can’t.

It’s only when the real physical world is brighter and more colorful and full of surprises. And luckily, I found something that holds my attention more than phones do. I found a certain set of things that —

Psychedelics.

[LAUGHS] Well, certainly psychedelics. Going out dancing, being face to face with a friend, reading, writing, right? Listening to music. Actually, this extremely limited set of things that are more mesmerizing to me and more pleasurable to me than the screen. And that’s like one of my very few concrete hopes that I have for my kids, is that they find something that makes the world’s dimensions enlarge in a way that overmatches however the world enlarges or seems to enlarge through a screen.

I want to get at how you’re getting there, because I think there’s something very deep in this question between attending to pleasure or attending to some joyful or meaningful dimension of experience, and attending to some of these other ends.

You talk in the piece about a researcher, and she casts no aspersions on the work she does, which sounds both necessary and annoying, who has done all this research trying to rate these different children’s shows according to how educational they are.

And so you have “Daniel Tiger,” and that gets a two. It’s a very weird 0 to 2 scale. But you have something like “Daniel Tiger,” which is an offshoot of the “Mister Rogers” Cinematic Universe, which gets a 2. Or maybe “Bluey” would probably get a 2. You have some “CoComelon,” which gets a 1. Not very educational, but not actively meant to be harmful and frightening. And then you have the strange underbelly of YouTube Kids, this sort of computer-generated, A.I.-animated, often kind of horrifying Dream Logic crap CGI, and that’s a zero.

And I was thinking, reading that, about what is implicit in this scale, which is that the best thing should be the most educational thing. And I don’t really believe that for myself. I’m not sure I believe it for children. But I think there is this question we face of, well, what is our measure? But what, for you, helps orient the tuning fork, both like for Jia or for your children?

I think within this question and this calculation, whatever I did in my own life and whatever part of that calculation led me to the idea that having kids is going to be a part of this pleasure-seeking — and it was indicative that the idea of pleasure had shifted a little for me, right? It wasn’t the pure hedonism of 12 years earlier or whatever, you know? It was something that was deeper and harder and more prolonged.

And I think part of it is, like, when I think about wanting my children to be oriented around pleasure, and that’s what my idea of a good life for them entails, it also involves them learning to conceive of pleasure as the things in life that make them feel more human. I guess, maybe that’s one of the ways I’ve clarified it for myself, right? The things that bring me pleasure are the things that make me feel more human and not less.

I think it’s interesting to say maybe we should be searching for pleasure as opposed to achievement. But that that also, there’s a lot of things one might want, right? Pleasure is one of them. But I mean, achievement is a reasonable one, too. And then there are all these —

If you find pleasure in it, right?

Right or maybe not, right? Maybe pleasure isn’t the point of life, or maybe — I mean, there’s all kinds of things that I do that I think are important or that I think are socially useful that I genuinely don’t find pleasurable.

But you don’t find a kind of hard pleasure in them, like a hard, sort of durational — you know what I mean? Do you ever like these things that — have you not fooled yourself into — [LAUGHS]

Sometimes somebody in the middle of a podcast will just press on a point in me so sore that I have to spend a minute to be like, am I really going to go here?

I’ll say, I guess, this. I am struggling with this question quite a lot lately in my own life where I am so driven, sometimes, by internal pressure that things that I think are pleasurable have been drained of their pleasure.

And literally, this morning, when I was getting ready for my day, I just have this note in my notebook about the things I need to do today. I’m like, can you try to be driven by something other than this internal pressure?

When I began writing about politics, I was a blogger in college before blogs were basically even a thing. There was no thought of it being a career. It was done for nothing but a kind of pleasure, right? A kind of delight in being engaged in the world and trying, in some small way, to understand it, and even in some even smaller, completely inconsequential way to influence it.

And now that I have this much bigger platform, it’s so much less pleasurable. And so, I think it’s interesting, this idea you’re getting at of expanding pleasure.

I think, well, back to the thing that you were talking about, where the researcher was coding stuff about what was educational or not, and that was the unspoken good, right, I think it’s definitely not coincidental that the things that were coded as maximally educational are also the things that parents find pleasurable to have on in the background, and that it’s not coincidental that the stuff that was not educational is videos where Minnie Mouse’s head falls off and rolls down a mountain, and that the middle ground was like this sort of “CoComelon,” “Blippi” thing that I think there’s something about kids probably can learn more when they are experiencing some sort of delight.

I hate “Blippi.” I just find “Blippi” completely unnerving. If you have never watched “Blippi,” go enjoy yourself on YouTube. But my kids like it.

Whoa! And look at what you rode up on — a police bicycle! Can I look at it?

Of course, you can, Blippi.

Oh, cool. OK, now look at this — a helmet. Wow. This keeps you nice and safe, OK? And ooh! Look up here. Do you see that? It’s a light. [IMITATING LASERS]

I do go back and forth a little bit. Maybe the reason they like it is that I don’t like it. They are different than me. They are not supposed to what I like. It gets to this broader sense of structuring everything around education and achievement. It’s like, we’re already thinking about this when they’re two?

Everything has to be educational. They can’t put on pants. They can’t put on shoes. And already, we are stretching out the tarp of their consciousness across a scaffolding of the adult world, right? Achievement, and education, and are you bettering yourself, and are you improving.

And I both get it, but it does feel like a transmutation of what we have done to ourselves onto them at a younger and younger and younger age, just this movement of a kind of what was once an elite adult culture, right? The sort of self-improvement culture, Dale Carnegie culture, lifelong education, to now it’s, like, your babies are supposed to be doing it. It feels odd.

Well, and I also think, to me, it disturbs me less because it is indicative of this broader culture of optimization that I abhor and participate in and find really pernicious. And I’m terrified of how it might advance itself upon my kids. I’m afraid of that.

But it also — nothing’s educational at this age anyway. Most of the “CoComelon” audience is not learning shit. I was wading through a lot of literature and research about this. Tiny little babies can process TV more than we think they can, but they can’t really learn anything anyway.

And I think that absolutely, it feels kind of overtly like a veneer that everyone is just pretending that we can talk about what is going to be good for them and what is healthy and what is not, while completely avoiding just there’s a big giant spotlight on what’s on the tablet and the whole world, the whole world, and all of the ways that the actual world will change the trajectory of their lives is kind out of focus.

The child psychologist Alison Gopnik — and I probably wouldn’t have brought this into conversation except that we’ve already been circling psychedelics a couple of times — has made this point. She’s at UC Berkeley. There’s been a lot of psychedelic research there. And so there’s been this interesting cross-pollination in those departments. And she’s made this point that the child’s brain looks a lot like the brain of an adult on psychedelics.

So it really does!

It really does. You have a lot more disorganization in the way the neurons are connecting a lot more. We learn, as we get older, to filter the world, right? And that’s not just a conceptual skill. That’s actually how our brains are organized.

Psychedelics disorganizes the brain, which is why people make a lot of unusual connections, and they’re absorbing an overwhelming amount of experience because they’re not filtering it out. There are other ways to get there, too. I remember when I came back from a silent meditation retreat, I was so unable to filter out visual information that I felt like I wasn’t safe to drive because just trees were too overwhelming.

But what was making me think about — the reason I bring it up here is that both in my own experience and people I’ve known, it’s like people, when they’ve had a psychedelic experience, when they turn on the TV at the end of it to kind of come to rest, if they decide to do that, they tend to watch cartoons. They watch Pixar. They don’t go for thoughtful adult movies.

And I think there’s some interesting analogy to that in this conversation about a children’s show, is, like, if a child’s brain is more psychedelic, more disorganized, more open, then in the same way that adults who have gone through those experiences want something more colorful, beautiful, safe, et cetera, that their orientation may be in that direction, too.

Maybe there’s something valuable in it, right? At the end of that experience, I don’t want something highly educational. And in the experience a two-year-old is having, wide-eyed in this completely overwhelming world, maybe they don’t and shouldn’t.

I think that’s why when it comes down to pleasure, I’m also — I once did an iconic three-movie comedown stream of — I think it was like “Ponyo,” then “Pocahontas,” then “Bambi” or something, you know? And I was like, this is living. This is pre-kids. But I think that the idea that anything kids watch should be one thing or another, I think we’re both kind of in disagreement. They’re kids. Let something just exist without a purpose maybe for a little bit.

But I also, for this reason, I think that just basic ideas of beauty and pleasure, they’re not that different from kid to adult. I mean, obviously, there are limits to this, right? Like, I was thinking about having a pacifier in your mouth all day long or whatever, but I think little children find the same things beautiful. We experience this every day. They are stunned by a leaf, a beautiful flower, looking at an animal, a picture of an animal, thinking about a whale. They’re oriented towards these things that we get to most readily in the psychedelic zone.

But this is maybe an argument for kids’ TV. We can maybe want it to just be beautiful. And I think we can want them to have an experience of beauty in a way that is not instrumentalized and has nothing to do with achievement.

If you’re going to be in front of a screen escaping or looking for something or just zoning out, why not have it just be legitimately delightful? And maybe “Blippi” is that for some kids.

But I came out of thinking about “CoComelon” for months with that idea in my mind. I was like, I think now that it might be a legitimate thing for me to want them to be looking at beautiful, stupid cartoons, the same ones that I would want to watch coming down from a hallucinogen.

There’s a way in which you manage what you measure. And I think the main way we have been taught to measure this or think about this question is, it’s always called the screen time question, right? The question of screen time. Do your kids have screen time yet? How much screen time? And something it feels to me like you’re getting at is that that’s just maybe the wrong way to think about this entirely.

I feel completely unbothered by, quote unquote, “screen time” when I am there with my kids. If we’re watching “The Incredibles” together, I do not think that is any less good of an experience for them, for almost any definition of the word good, than if we go to Target together or if we’re just having fun together. Like, that’s a good experience.

Do you think that we have just sort of lost the plot on this sort of altogether, maybe for kids and adults and sort of making this about almost like the existence of the screen, rather than the experience of the person?

Yeah, the experience of the person. I mean, right. So if you actually get into the studies on the actual effects of screen time, it is like the screen itself is almost a red herring, right? People think about screen time as it correlates with achievement and verbal abilities and self-regulation and language abilities at grade level 7, all these things that are tracked longitudinally.

And the correlation is much more strongly between the kind of life the kid has and those things, right? We all kind of sense that it’s not — the screen is not the singular determining factor. It’s just that we put screens in use in ways that reflect the life of the child holistically, and the kinds of opportunities they have, and the kind of household they’re raised in, and the freedom that they have to not be thinking about basic needs, and to flourish in these other realms that we call achievement.

There are researchers that argue that children’s screen time use should be reframed as an indicator of parental distress. You know what I mean? It’s the life that matters, I think. And I think that applies to us and our smartphones, too. Like, when I was kind of required for one job or another to be constantly paying attention to the news as it was scrolled out on Twitter all day long, I was glued to social media kind of by requirement.

And I think the way I thought of it then was like, this is bad, but it’s OK as long as my real life is bigger than it. As long as I self-evidently always feel that the physical world is more inviting to me than my screen, then I’m not going to spend one second worrying about my brain rot because there’s nothing I can do about it, you know? I think as long as the world is winning out most of the time, I think that’s a reasonable — it feels like a reasonable metric to me.

I think so much of why parents hate “CoComelon” is a kind of self-loathing, often born of a kind of fatalism. It’s like, I don’t like this, but I’m doing it anyway, because other parents do it, because I need it, because I can’t think of an alternative, because I don’t have the energy to structure things differently, which is I’m not saying this is a thing true for other parents and not me. This is a thing true for me.

And I think that one reason I’m so interested in this conversation is that the conversation for kids feels, not exactly, but in many ways, like a miniature and clarified version of the conversation for adults. And weirdly, we’re better at having the conversation for kids because at least there, we can imagine making judgments about good and bad and about using paternalism and cultural pressure.

Not that many people I know are truly happy with their digital lives if they’re in a constant state of irritation and aggravation with themselves above all. But there’s something about the fact that everybody else is there or feels like everybody else is there that makes it impossible to imagine or effectuate a different reality, even just for yourself, even when such a reality is possible.

Mm-hmm. I also think there’s something about what you’re saying where you do something on your phone, and it’s unsatisfying. And you’re dissatisfied with the way that you’re doing it. But the smartphone has become the repository for all possible dissatisfaction and yearning.

Civic dissatisfaction — routed through the smartphone. Social — you feel lonely, you go to the thing that’s making it worse, right? I do think that we’re kind of in the grip of the loop where the dissatisfaction with the thing itself presents it as the answer, right? If you lack money, be a TaskRabbit, or drive for Lyft, or deliver. There’s a way that the phone is the catchall solution for any sort of discontent.

Let’s say someone is still using Twitter, and they’re miserable, and they want to get off of it. Probably they’re still going to be looking for a source, a replacement that exists on the phone, you know? OK, maybe not that one, but what about Reddit, or something like that?

What I think is so hard about that is that what makes it impossible to have alternatives to a bad status quo is the continued investment in the status quo. But because it maintains just enough staying power, that is energy people are not being able to put in to creating things that are different. And I think there are things that are different out there or certainly things that are different that could be imagined.

And again, I think this ends up being true for kids, too. We are very social creatures. And what everybody else does really ends up mattering. I mean, kids see other kids in a restaurant, and the other kids are allowed to watch a phone during the meal. And that makes it harder to resist your kids wanting to do that, right? I mean, there’s this whole pressure about getting kids smartphones in school because their friends have smartphones.

And again, there is something so contagious about everything, and there’s something that is so true in the way that the existence of something fairly totalizing or fairly central, the fact that you have to participate in it, particularly if it’s something that drains your attention and creative energy, the fact that you have to participate in it at all, or certainly, if you have to participate in it a lot, it makes it that much harder for other things to emerge, because they would need to emerge in that same space with that same energy.

I had television growing up. I don’t think it was terrible for me, but I do wonder about how different it is that my kids can watch anything at any time, whereas I was at least a little bit prisoner to what was on when, including, I had it easier. I had Nickelodeon, which had a lot of kids’ programming, right? I didn’t only have to watch it on the kids’ hour on network television. With all these things, it feels like there’s some balance that makes sense, right? Some point where it’s enough and not too much. Enough escape, but not too much escape. Enough choice, but not too much choice. And there’s just a lot of things it feels like we’ve, to me, hit too much. And maybe that’s just me getting old. And I just think when I was a kid, it was enough, and now it’s too much. But I also think there has to be conceptually too much has to be a possibility, and maybe we’ve reached it.

Well, I think, again, I mean, it’s like my dumb ass keeps bringing this all back to pleasure. But I think that feels like that line, right? Like, once you’d had enough of watching six episodes of “Pete and Pete” in a row, which I certainly did, once you were feeling severely diminishing returns, you would walk away from this machine that was not watching you and was not altering its behavior to get you to watch it more.

And we were able to do that. We were able to have this kind of unadulterated, physical, cognitive instinct about what was the right amount of escape and what was the right amount of engagement, because we just followed what we wanted to do. We weren’t walking away from six hours of “Pete and Pete” because it was good for us. It would be better for our language abilities in the seventh grade if we did so.

We were just like, I’m bored. I’m no longer getting pleasure from this. I’ve actually not been getting that much pleasure out of the last two episodes. I’m just going to go do something else for a while.

I feel like every time I write about children’s or anything, any sort of media, it’s like everyone has been having these exact same worries. I was even thinking, with this conversation, you know the concept of acedia?

OK, so this is like, there’s a joke. Me and my friends have a joke that it would be a beautiful name for a girl, where it was this medieval conception of depression that, to me, feels like exactly like what we talk about when we talk about smartphones.

I looked it up. I looked it up on Wikipedia last night to make sure that I was — and there was this beautiful description on Wikipedia of acedia as a flight from the divine that leads to not even caring that one does not care. It’s like this listlessness, this disengagement from the world, this boredom. And then you start not even caring that you’re so bored, like this total inability to act upon your life.

As with the example of we were able to walk away when we weren’t having fun anymore, we didn’t have the option of the TV just being like, wait, wait, wait. Try these 45 other things. I’ll hypnotize you again in 45 seconds if you just give me the chance, right? I can feel it interfering with my own ability to understand when something’s fun and when it’s not.

And I think that I am worried about that with my children. And I think that’s one of the reasons that I’m like, the only way out is a set of experiences or desires that will clean out and clarify your radar for what actually feels good and what actually doesn’t and, yeah, feels good in any of the ways — the meaningful ones, the not meaningful ones.

But it seems to me like one of the things that that responsive surveillance mechanism does is it mixes all of those things up, so that even if you’re no longer having fun on Twitter, there’s still some part of you that feels like you are, just because of the mechanism itself.

I think this is a place where the surveilled language and fear really does hit for me. Surveillance, it sounds so creepy, and it is. And it’s part of it. But I actually feel like it masks the reason we give ourselves over to it because it feels good to be learned about.

We all have experiences of the algorithms coming to know us or predict us and recommending a book we’d never read before that actually was really great, or music that we never heard before that brought us into a whole new genre or a whole new artist that we would have never found on the radio necessarily, or tweets that we are glad we saw.

But it’s that way of being learned so it is able to continuously recommend things that are more and more alluring. And it makes that experience of the diminishing marginal return more distant, right? That place where it’s like, well, I’ve already watched two episodes of “Pete and Pete.” I don’t want to watch three. That’s a lot of episodes of “Pete and Pete.” Instead, it knows better what you want and is better at giving it to you. It’s why I find I really try never to leave my kids alone with a recommendation algorithm, right, YouTube or anything like that. It’s much scarier. And where they end up is much worse. But it is learning them at every second and how to give them the thing they actually want that will keep them clicking.

And again, I think this is where “CoComelon” feels weird. I mean, “CoComelon,” to bring it back to that, it’s one of the first successes in children’s television or video entertainment, I guess you’d call it, that doesn’t come out of television. It comes out of YouTube. It comes out of recommendation algorithms. It’s built around recommendation algorithms.

And I feel like it feels like recommendation algorithms, right? It feels like it knows the kids too well. It feels too tuned to their short attention spans. The whole thing is just overly optimized. And on one level, that makes it very effective as a babysitter or an attentional harvester, or just maybe it’s better as entertainment for them.

But it creates, in this very clear way because you’re watching it happen to a two-year-old, this feeling of what it looks like when culture is built because it knows you, and it knows how to predict you. And that scuzzy feeling that we have in a vaguer way, I think, with ourselves when we’re older, we get into this very intense way with them when they’re younger. But it’s the same thing, in a way, to me, like all up and down the age ladder.

I also think that, again, this is a thing where the thing that you get through the smartphone, like you were saying, this experience of being learned and being known very, very deeply, it’s a really human desire that has made this so effective as an addictive technology. Of course, we want to be learned, and we want to be known. That’s so much of the entire pleasure of being alive interpersonally.

We want that for good reasons. But it’s like, yeah, it’s part of why “CoComelon” and so much on the phone, it feels frictionless. It’s been designed for frictionlessness. And I think that’s part of — I mean, maybe this is helping me understand how I delineate good pleasure, meaningful pleasure, from meaningless pleasure, which is that I think there’s friction in all real pleasure and in the kind of pleasure you learn to get in the real world. There’s friction in it. There’s true surprise.

And I think that when someone is learning us in the real world, when they’re coming to know what we would like and they’re seeing things about us that we don’t even see, all of these things that the algorithm is doing, when another person is doing that for us, they change us in ways that the algorithm doesn’t, right? That experience contains sharp edges in a way that the algorithmic one never can.

And that’s why it’s almost the same thing, but it’s why the real world version, it feels infinitely more meaningful than Spotify learning what song I want to listen to next. Because even as good as the algorithm genuinely has gotten at giving me what I want to listen to, the total removal of friction — and I think that’s one of the reasons that all the YouTube stuff feels just instinctively bad — again, it’s one of those things where maybe it’s easier to see with “CoComelon.”

It’s easier to see that the seamless microtargeting, endless stream of giving you exactly what you want, I’m certainly guilty of understanding that it’s good for my kids to have a little more friction in their life and not just get everything what they want at a touch of a button and then me pursuing that as soon as they’re in bed.

This maybe gets at another thing that I’m afraid of, because one of the things I was thinking about while you were saying that is, I decided a couple of months ago, I would sign up for a bunch of the A.I. relationship apps, right, like Kindroid and Character AI and things like that, where these language models are built to create a kind of A.I. you’d be friends with or a lover with, or it would be your therapist or whatever.

And I tried them out for a while. And they’re pretty good now. I mean, they’re good at texting. What they write sounds and feels realistic. I always tell people they’re much better at texting than most of the people I know.

Would it have fooled you?

It doesn’t fool me. Yeah, it would definitely have fooled me.

It would have successfully catfished you if one of them had started texting you?

100 percent, easily. I would not have known it’s not a person. But I never could keep myself coming back to them because there was no meaning in the interaction, right? So since moving to New York, I’ve been making new friends.

And I was thinking about how one of the friends I made, we text a lot just sort of during the day. And they’re not interesting texts, necessarily. Sometimes they are, but it is meaningful to me that he is giving me that attention back, right? The message of the text is that I am being chosen for somebody else’s attention. It’s a kind of meta text about a relationship that is emerging.

And I didn’t end up writing a piece on this, though, because I can see the Character AI usage numbers that are being released. And this is a sort of A.I. system used much more by younger people. And they’re logging in a ton of times a day and spending huge amounts of time on it.

So, to go to the point of what I’m afraid of, right, this question of the retraining, for me, who grew up before large language models, the kind of uncanniness, what I would call the meaninglessness, of that interaction is very front and center, right? It’s very noticeable.

But if you’re younger, and your social dynamics are way less formed, and your discernment of social dynamics is much less mature, and your choices are more limited because of who you know and how you can see them and how you can be in touch with them, maybe it actually doesn’t feel that way, or maybe you get trained out of it feeling that way. I do think you can lose the sense of and taste for friction, and that that is a loss and that it does foreclose forms of pleasure, just as you were talking about.

Hugely. There’s this old Kurt Vonnegut thing where he was talking about the pleasure of mailing a letter. And it’s like how the whole point of it is not that you’re doing something efficient. The whole point of it is that you go for a walk, and you wink at a girl, and you pet a dog or whatever.

I mean, but yeah, I get used to — I used to — pre-children, I had a years’ long streak of never using Amazon and a personal policy of, if it was within walking distance of me in downtown-adjacent Brooklyn, then I could not order it online. I had to physically go out and get it. And I have backslid on all of this since having my second child significantly.

And it’s also partly because the experience of having children, wonderfully, it is the source of all that friction. It’s sort of like back to what we were talking about at the beginning. I think that as our world orients itself increasingly towards frictionlessness, children can seem exclusively like a form of friction. And that friction can seem exclusively like something that’s undesirable, when, in fact, I think my sense that I wanted some of this specific kind was one of the things that made me think it would be fun.

And I do think that a total lack of impediment, the ability to pick up and go anywhere you want at the drop of a hat, which is a wonderful way to live, which anyone who has it and wants it, I’m jealous of you, and good for you. I mean, it’s like everything that we’ve been talking about is, those are values that have been inculcated by the same form of ultra-advanced capitalism that created the smartphone and created all these things that make us so depressed in the first place.

And I think that part of me wanting, feeling ready to try to have kids four years ago, five years ago, whatever it was, was a sense that I wanted to undo these things in me, like this sense of exceptionalism, which children really took out of me, you know? We were texting about the thing where you’re playing with your kids. And you’re like, I’m probably not particularly good at being the horsey right now, but the thing that matters to you is that I am the person being the horsey.

The most common way I feel like I fail as a parent — and it’s the way I fail both my kids and myself — is by trying to really control and optimize the experience and treat it like other things in my life. We’re going to do this at this time, and then we’re going to here. And you got to get your shoes on by this moment.

And to your point about being the horsey or finding some joy in submission, parenting is so unpleasant when you feel like it is a distraction from the thing you would prefer to be doing, right, like looking at your phone or taking a nap. But it’s also, I think, unpleasant when you are trying to treat it like other things in adulthood and control it. It’s most pleasant for me when I have the resources inside myself and also the wisdom to just kind of be around. They’re running around and occasionally playing with me. And I’m sitting on the couch and occasionally playing with them. And it’s like you could just do a lot less. We’ve made parenting really, really hard. And we put a lot of pressure on ourselves as parents to try to do a great job of it and be achievement oriented. And the kids should be watching only educational shows, when, really, they should be watching no shows at all.

And there’s a million things we’ve done that are not really anything that the kids ever asked us to do. They would like you to be around a bit more or a lot and be attentive, but also not overtake their experience with your own. And that’s really hard. I would like to get better at that.

I think that people know this, though. I think that when I had my first kid, the thing that I found most difficult, but now I think I’ve smooth-brained my way into finding really pleasurable is that when you’re with your kids, at some point, you just have to completely surrender to not — the time just really can’t — you can’t really do anything else, you know? You are just going to — this is your weekend now. This is what weekends are going to be like now.

And there’s a removal of choice in that that is the thing that I was afraid of and I think a lot of people are afraid of, but also the thing that is arguably the most freeing. I remember also, I had read “How to Do Nothing” right around. And I was like, yeah, this might be a shortcut to the kind of outside the clock time, where you are just not being useful to anyone but the people in front of you, this thing that I was trying so hard to do in other ways that now I have to do every single weekend, whether I like it or not.

And I no longer feel that as a loss, I’m realizing. I used to. I used to think, oh, I could have done so many things with these weekends. And now I’m like, oh, it’s time to go to the playground, you know. Time to go to the playground again.

We’ve ranged a lot here. But to go back to, in some ways, the article that led to this show, if a kid watches an hour of “CoComelon” a day, should you feel bad? Would you feel bad?

No. Should I go on? [LAUGHS] No, I don’t think so.

I think that’s a lovely place to end. So always our final question, what are three books you would recommend to the audience?

I forgot about this question until I was on the train here, and I was like, OK, I got to think about the last three things that I was just texting people about because I really loved. OK, so I was extremely late to becoming “Lonesome Dove” pilled. I was like six months pregnant, and I was on a work trip to Thailand. And I had 48 hours of alone time on the end of that work trip. And I was like, this is the most important —

the books I bring are the most important books. This is the only time I’ll be alone for 48 hours all year, you know? And I was like, I need to bring the perfect book, the book that will make me feel like I’m a kid again, will give me just this wildly disproportionate, emotional attachment. I want to be sobbing by the end of this book. And I read “Lonesome Dove,” and it was all that and more. I have been “Lonesome Dove” pilling many of my friends as the year has gone on. If anyone hasn’t read it, truly recommend it.

I really like this book, “In Ascension,” that our friend Max recommended to me. If anyone is in the Ted Chiang, Jeff Vandermeer kind of thing, it’s like that kind of grounded, beautiful, enigmatic, slightly schematic sci-fi. Really loved it, another one that I’ve been texting a lot of people about.

The third one is “When We Cease to Understand the World” by Benjamin Labatut, one that I feel like I’ve been texting friends about every month since I read it. It’s about scientific discoveries that bring people to the brink of madness. And there’s a really interesting thing that goes on where the book starts off almost entirely nonfiction and then ends almost entirely fiction. And the gradations in between are amazing.

And I’ll just note, because you mentioned in your last answer, you talked about reading “How to Do Nothing,” which is by Jenny Odell. And I still think it’s the best book about attention and habits of mind —

Incredible.

— in this era. I’ve enjoyed this so much, Jia Tolentino. Thank you very much.

Thank you. [MUSIC PLAYING]

This episode of “The Ezra Klein Show” was produced by Annie Galvin, fact-checking by Michelle Harris with Mary Marge Locker. Our senior engineer is Jeff Geld, with additional mixing by Isaac Jones and Amon Sahota. Our senior editor is Claire Gordon.

The show’s production team also includes Rollin Hu, Elias Isquith and Kristin Lin. We have original music by Isaac Jones. Audience strategy by Kristina Samulewski and Shannon Busta. The executive producer of New York Times Opinion Audio is Annie-Rose Strasser. And special thanks to Sonia Herrero.

Produced by ‘The Ezra Klein Show’

I feel that there’s something important missing in our debate over screen time and kids — and even screen time and adults. In the realm of kids and teenagers, there’s so much focus on what studies show or don’t show: How does screen time affect school grades and behavior? Does it carry an increased risk of anxiety or depression?

And while the debate over those questions rages on, a feeling has kept nagging me. What if the problem with screen time isn’t something we can measure?

[You can listen to this episode of “The Ezra Klein Show” on the NYT Audio App , Apple , Spotify , Amazon Music , YouTube , iHeartRadio or wherever you get your podcasts .]

In June, Jia Tolentino published a great piece in The New Yorker about the blockbuster children’s YouTube channel CoComelon, which seemed as if it was wrestling with the same question. So I invited her on the show, and our conversation ended up going places I never expected. Among other things, we talk about how the decision to have kids relates to doing psychedelics, what kinds of pleasure to seek if you want a good life and how much the debate over screen time and kids might just be adults projecting our own discomfort with our own screen time.

We recorded this episode a few days before the Trump-Biden debate — and before Donald Trump chose JD Vance as his running mate. We then got so swept up in politics coverage we never got a chance to air it. But I am so excited to finally get this one out into the world.

You can listen to our whole conversation by following “The Ezra Klein Show” on the NYT Audio App , Apple , Spotify , Amazon Music , YouTube , iHeartRadio or wherever you get your podcasts . View a list of book recommendations from our guests here .

(A full transcript of this episode is available here .)

This episode of “The Ezra Klein Show” was produced by Annie Galvin. Fact-checking by Michelle Harris, with Kate Sinclair and Mary Marge Locker. Mixing by Isaac Jones, with Efim Shapiro and Aman Sahota. Our senior editor is Claire Gordon. The show’s production team also includes Jeff Geld, Rollin Hu, Elias Isquith and Kristin Lin. Original music by Isaac Jones. Audience strategy by Kristina Samulewski and Shannon Busta. The executive producer of New York Times Opinion Audio is Annie-Rose Strasser. Special thanks to Sonia Herrero.

Follow the New York Times Opinion section on Facebook , Instagram , TikTok , WhatsApp , X and Threads .

IMAGES

3. FIND S Algorithm Finding a Maximally Specific Hypothesis in Machine Learning
Lec-15
Hypothesis in Machine Learning
Everything you need to know about Hypothesis Testing in Machine
4. FIND-S Algorithm Finding A Maximally Specific Hypothesis Solved Example 2 in Machine Learning
FIND S Algorithm

VIDEO

Hypothesis Testing in Machine Learning
Inferential Statistics
Probabilistic ML
Probabilistic ML
Unit-1 Machine Learning
Hypothesis spaces, Inductive bias, Generalization, Bias variance trade-off in tamil -AL3451 #ML

COMMENTS

Finding a Maximally Specific Hypothesis: Find-S
Finding a Maximally Specific Hypothesis: Find-S . The find-S algorithm is a machine learning concept learning algorithm. The find-S technique identifies the hypothesis that best matches all of the positive cases. In this blog, we'll discuss the algorithm and some examples of Find-S: an algorithm to find a maximally specific hypothesis.
ML
ML | Find S Algorithm. The find-S algorithm is a basic concept learning algorithm in machine learning. The find-S algorithm finds the most specific hypothesis that fits all the positive examples. We have to note here that the algorithm considers only those positive training example. The find-S algorithm starts with the most specific hypothesis ...
FIND S Algorithm
3. Apply the FIND-S algorithm by hand on the given training set. Consider the examples in the specified order and write down your hypothesis each time after observing an example. Step 1: h0 = (ø, ø, ø, ø, ø) Step 2: X1 = (some, small, no, expensive, many) - No. Negative Example Hence Ignore. h1 = (ø, ø, ø, ø, ø)
Find S Algorithm in Machine Learning
The S algorithm, also known as the Find-S algorithm, is a machine learning algorithm that seeks to find a maximally specific hypothesis based on labeled training data. It starts with the most specific hypothesis and generalizes it by incorporating positive examples. It ignores negative examples during the learning process.
Machine Learning- Finding a Maximally Specific Hypothesis: The List
Specific Hypothesis: If a hypothesis, h, covers none of the negative cases and there is no other hypothesis, h′, that covers none of the negative examples, then h is strictly more general than h′, then h is said to be the most specific hypothesis. The specific hypothesis fills in important details about the variables given in the hypothesis.
Discover Power of Find S Algorithm
Return the maximally specific hypothesis, if it exists. The Specific-to-General Algorithm is a simple and efficient way to learn a maximally specific hypothesis from a set of training examples. It is widely used in machine learning and has applications in many different domains, such as natural language processing, computer vision, and robotics.
FIND S Algorithm
FIND S Algorithm | Finding A Maximally Specific Hypothesis | Solved Example - 1 by Mahesh HuddarSolved Example 2 - https://www.youtube.com/watch?v=SD6MQLC2Dd...
Find-S Algorithm In Machine Learning: Concept Learning
In Machine Learning, concept learning can be termed as "a problem of searching through a predefined space of potential hypothesis for the hypothesis that best fits the training examples" - Tom Mitchell. In this article, we will go through one such concept learning algorithm known as the Find-S algorithm. If you want to go beyond this article and really want the level of expertise in you ...
Version space learning
GB is the maximally general positive hypothesis boundary, and SB is the maximally specific positive hypothesis boundary. The intermediate (thin) rectangles represent the hypotheses in the version space. Version space learning is a logical approach to machine learning, specifically binary classification.
Hypothesis in Machine Learning
A hypothesis is a function that best describes the target in supervised machine learning. The hypothesis that an algorithm would come up depends upon the data and also depends upon the restrictions and bias that we have imposed on the data. The Hypothesis can be calculated as: y = mx + b y =mx+b. Where, y = range. m = slope of the lines.
Concept Learning: The stepping stone towards Machine Learning with Find
The goal of a learner is to find a hypothesis h which can identify all the objects in X so that: h (x) = c (x) for all x in X. In this way there are three necessary things for an algorithm which supports concept learning: 1. Training data (Past experiences to train our models) 2. Target Concept (Hypothesis to identify data objects) 3.
Chapter 2: Concept Learning and the General-to-Specific Ordering
The specific boundary S, with respect to hypothesis space H and training data D, is the set of maximally specific members of H consistent with D. Version Space Representation Let X be an arbitrary set of instances and let H be a set of boolean-valued hypotheses defined over X.
Maximally Specific Hypothesis
Maximally Specific Hypothesis. Find-S Algorithm Concept Learning Hypothesis Partial Ordering. Maximally Specific Hypothesis. Begin with the most specific possible hypothesis in , generalize this hypothesis each time it fails to cover an observed positive training example. Find-S algorithm ignores negative examples. If the hypothesis space ...
Machine Learning- General-To-Specific Ordering of Hypothesis
Reference. General-To-Specific Ordering of Hypothesis. The theories can be sorted from the most specific to the most general. This will allow the machine learning algorithm to thoroughly investigate the hypothesis space without having to enumerate each and every hypothesis in it, which is impossible when the hypothesis space is infinitely vast.
FIND S Algorithm in Python
Python Program to Implement FIND S Algorithm - to get Maximally Specific Hypothesis. Exp. No. 1. Implement and demonstrate the FIND-S algorithm in Python for finding the most specific hypothesis based on a given set of training data samples. Read the training data from a .CSV file. Find-S Algorithm Machine Learning 1.
4. FIND-S Algorithm Finding A Maximally Specific Hypothesis Solved
4. FIND-S Algorithm Finding a Maximally Specific Hypothesis Solved Example 2 in Machine LearningFollowing concepts are discussed in the video:*****...
PDF Version Space Learning
CSG220: Machine Learning Version Space Learning: Slide 2 Concept Learning • Recall that a 2-class classification problem is also called a concept learningproblem • Given the mapping f : X Æ{+, -} to be learned, the set of instances f-1(+) X is the corresponding concept • A concept is just a subset of the instance space X,
What is a Hypothesis in Machine Learning?
A hypothesis is an explanation for something. It is a provisional idea, an educated guess that requires some evaluation. A good hypothesis is testable; it can be either true or false. In science, a hypothesis must be falsifiable, meaning that there exists a test whose outcome could mean that the hypothesis is not true.
Find-S Algorithm: Finding maximally specific hypotheses
The search begins (ho) with the most specific hypothesis in H, then considers increasingly general hypotheses (hl through h4) as mandated by the training examples. ... There can be several maximally specific hypotheses consistent with the data. Find S finds only one; ... Machine Learning More. Advertisement NumPY Tutorial Pandas Matplotlib ...
ML
Introduction : The find-S algorithm is a basic concept learning algorithm in machine learning. The find-S algorithm finds the most specific hypothesis that fits all the positive examples. We have to note here that the algorithm considers only those positive training example. The find-S algorithm starts with the most specific hypothesis and generali
machine learning
If there are several maximally specific hypotheses that fit a data set, Find-S will just return one of them, where as C-E will return all of them as part of the specific boundary of the version space. If there is only 1 maximally specific hypothesis though, there is no difference. Hope this helps!
Hypothesis in Machine Learning
The hypothesis is one of the commonly used concepts of statistics in Machine Learning. It is specifically used in Supervised Machine learning, where an ML model learns a function that best maps the input to corresponding outputs with the help of an available dataset. In supervised learning techniques, the main aim is to determine the possible ...
PDF What is MACHINE LEARNING?
Machine learning(ML) studies the construction and analysis of algorithms thatlearn from data. ... hypothesis space. A hypotheses is a function h : X! f 0;1g. The goal is to nd a hypothesis h such that h = c, a task that may be unattainable since we know the values of c only on S.
FlowRetrieval: Flow-Guided Data Retrieval for Few-Shot Imitation Learning
Few-shot imitation learning relies on only a small amount of task-specific demonstrations to efficiently adapt a policy for a given downstream tasks. Retrieval-based methods come with a promise of retrieving relevant past experiences to augment this target data when learning policies. However, existing data retrieval methods fall under two extremes: they either rely on the existence of exact ...
Decoding emotions: unveiling the potential of facial landmarks
Therefore, the data utilized for both training and testing in machine learning comprised behavioral measurements, facial landmark data, and a combination of both. Data processing. We streamlined the datasets related to facial expressions through a comprehensive data processing phase, considering the abundance of information and landmark features.
JPM
Recent scientific research has shown that the ketogenic diet may have potential benefits in a variety of medical fields, which has led to the diet receiving a substantial amount of attention. Clinical and experimental research on brain tumors has shown that the ketogenic diet has a satisfactory safety profile. This safety profile has been established in a variety of applications, including the ...
Opinion
transcript. On Children, Meaning, Media and Psychedelics The writer Jia Tolentino on parenting — and living a good life — in the age of smartphones.

Find S Algorithm in Machine Learning

What is the Find-S algorithm in Machine Learning?

Symbols used in Find-S algorithm

Inner working of Find-S algorithm

Kickstart Your Career

Discover the Power of Find S Algorithm – A Comprehensive Guide

What is find s algorithm in machine learning?

Find s algorithm in machine learning with an example

Find s Algorithm advantages and disadvantages.

Advantages: –

Disadvantages: –

Limitations of find s algorithm

Difference between find-s and candidate elimination algorithm

Why do we use the Find-S algorithm?

What is the output obtained by Find-S algorithm?

What is the algorithm for finding a maximally specific hypothesis?

How does find s algorithm start from the most specific hypothesis and generalize it?

RELATED ARTICLES MORE FROM AUTHOR

Project Management Skills You Need to Succeed in an IT Company

Your Guide to Starting a Small Business as a Solopreneur

Learn the Secrets of Winning on Twitter

Data Science Introduction

Statistical Inference

Machine Learning

Predicting the Outbreak of COVID-19 Pandemic using Machine Learning

Top 10 Applications of Machine Learning in Daily Life

Supervised Learning

Unsupervised Learning

Miscellaneous

Career Opportunities

Interview Questions

Artificial Intelligence

Machine Learning Full Course – Learn Machine Learning 10 Hours | Machine Learning Tutorial | Edureka

What is Find-S Algorithm in Machine Learning?

Limitations of Find-S Algorithm

How Does It Work?

Top 10 Trending Technologies to Learn in 2024 | Edureka

Recommended videos for you

ChatGPT Training Course: Beginners to Advance ...

Artificial Intelligence Certification Course

Prompt Engineering Course

Generative AI in Business: University of Camb ...

MLOps Certification Course Online

Large Language Models (LLMs) Certification Co ...

Reinforcement Learning

20,00,000 learners love us! Get personalised resources in your inbox.

Concept Learning: The stepping stone towards Machine Learning with Find-S

Share the Knol:

Written by Girish Bharti

6 thoughts on “ Concept Learning: The stepping stone towards Machine Learning with Find-S 9 min read ”

Discover more from Knoldus Blogs

Chapter 2: Concept Learning and the General-to-Specific Ordering

Inductive Learning Hypothesis

Find-S Algorithm

Candidate Elimination Algorithm

Version Space Representation

Candidate Elimination Algorithm Issues

Inductive Bias

Sample Exercise

Definition:

FIND S Algorithm in Python

Find-S Algorithm Machine Learning

Solved Numerical Example – Find S Algorithm to Find the Most Specific Hypothesis

Related Posts

Welcome to VTUPulse.com

C Programming

Machine Learning

Data Base Management

Data Structures

Saturday 16 July 2022

Unanswered questions by FIND-S

0 comments :

Advertisement

ML – Candidate Elimination Algorithm

Please Login to comment...

Improve your Coding Skills with Practice

What kind of Experience do you want to share?

Collectives™ on Stack Overflow

Is there any difference between most specific hypotheses obtained by Candidate Elimination and Find-S methods?

Your Answer