✦ScienexySCIENCE KNOWLEDGE LIBRARY
Science reference · SCI-CS-003-052

Definition and Scope of Machine Learning

Machine learning is the study of algorithms that improve their performance on a task through experience, with applications across prediction, classification, pattern recognition, and automated decision-making.

Learning = Improvement in Performance P on Task T through Experience E

Explanation

Machine learning can be formally described using a definition proposed by Tom M. Mitchell: a computer program is said to learn from experience E with respect to some class of tasks T and performance measure P if its performance at tasks in T, as measured by P, improves with experience E. This definition highlights three essential elements: the task being performed, the experience available to the system, and the measure used to evaluate improvement. The scope of machine learning includes supervised learning, unsupervised learning, semi-supervised learning, self-supervised learning, and reinforcement learning. Its methods are applied to classification, regression, clustering, dimensionality reduction, anomaly detection, recommendation, forecasting, and generative modeling. Machine learning also includes important supporting areas such as feature engineering, model selection, optimization, evaluation, interpretability, and responsible deployment. Modern ML systems can process structured data, images, audio, video, and text. However, machine learning is not automatically appropriate for every computational task. Some problems are better solved with ordinary algorithms or explicitly defined rules, especially when the requirements are deterministic and easily specified.

Variables

SymbolMeaning

T

The task or class of tasks a machine learning system performs.

E

The experience or data from which the system learns.

P

The performance measure used to evaluate improvement.

Example or application

For a handwriting recognition system, T is recognizing handwritten characters, E is a collection of labeled handwriting examples, and P may be the percentage of correctly recognized characters on a separate test dataset.

Text

The scope of machine learning extends beyond making predictions. It includes recognizing patterns, discovering hidden structures, generating content, recommending actions, and learning effective behavior from interaction.

Key Points

  • Mitchell's definition describes learning through improvement on tasks using experience.

  • Task, experience, and performance measure are central parts of the definition.

  • Machine learning covers predictive, descriptive, and interactive learning methods.

  • Its scope includes data processing, model development, evaluation, and deployment.

  • Not every problem requires machine learning.

Applications

  • Predictive analytics

  • Computer vision

  • Natural language processing

  • Speech recognition

  • Recommendation and personalization

  • Time-series forecasting

  • Generative AI

Learning Outcomes

  • Explain the formal definition of machine learning.

  • Identify task, experience, and performance measure in an ML example.

  • Describe major areas within the scope of machine learning.

  • Recognize situations where conventional programming may be more appropriate.

Practice Questions

  • How did Tom Mitchell define machine learning?

  • What do T, E, and P represent?

  • Give three examples of tasks solved using machine learning.

  • Is machine learning necessary for every software problem? Explain.

Related Concepts

  • Introduction to Machine Learning

  • Artificial Intelligence

  • Statistical Learning

  • Model Evaluation

  • Machine Learning Applications

Common Mistakes

  • Leaving out the performance measure when explaining Mitchell's definition.

  • Assuming machine learning is limited to prediction.

  • Confusing the scope of machine learning with only deep learning.

  • Assuming every program that processes data is a machine learning system.

Sources and verification

Sources are shown as provided in the project data. Not every source has yet been independently verified. Please report errors or missing citations.

  • Mitchell, T. M. Machine Learning, 1997.
  • Bishop, C. M. Pattern Recognition and Machine Learning.
  • Murphy, K. P. Probabilistic Machine Learning: An Introduction.

Suggest a correction or source · More in Artificial Intelligence & Machine Learning