Definition of machine learning
- Definition 1
- Arthur Samuel (1959). Machine Learning:field of study, gives computers the ability to learn without being explicitly programmed.
-
- The definition given by Arthur Samuel in 1959: Machine learning is a field of study that makes a computer capable of learning in the case of ambiguous programming,
- Definition 2 (not very easy to understand)
- Tom Mitchell (1998) well-posed Learning PROBLEM:A computer program was said to learn from experience E with respect to SOM E Task T and some performance measure p, if its performance on T, as measure by P, improves with experience E.
-
- Tom Mitchell in 1998: (Google Translate) Appropriate learning problem: is a computer program from experience E to learn a task T and some performance measurement p, if its performance on T, measured by P, with experience E and improve. (Later on as an example)
- #################################################################
- Example:suppose your email program watches which emails "do" or "do" mark as spam, and based on this learns how to B etter filter spam. What's the task T in this setting?
- Classifying emails as spam or not spam. ---T
- Watching you a label emails as spam or not spam. ---E
- The number (or fraction) of emails correctly classified as spam/not spam. ---P
- Example: Suppose your e-mail program observes your flagged spam and non-spam messages and learns how to better filter spam based on observations, what is T in this example?
- Classify messages as spam and non-spam---T
- Look at you. Mark e-mail messages as spam or non-spam---E
- The number of e-mail messages (or fractions) correctly categorized as spam/non-spam---P
- Example:playing Checkers.
- Example: Checkers Games
- E = play a lot of the experience from checkers
- T = The task of playing checkers
- P = The possibility of the program winning the game
Classification of machine learning
- Supervised learning (supervised learning)
- Unsupervised learning (unsupervised learning)
- Others:reinforcement Learning, Recommender systems. (Enhanced learning, referral system)
Why is it so popular now?
Machine learning
- Grew out of the work in AI (from the traditional artificial intelligence (can be understood as having its own set of systems))
- New capability for computers (computers can learn by themselves).
Machine Application Field
Examples:
- Database mining (Data mining)
- Large datasets from growth of automation/web. (Large data sets for automation/network growth)
- e.g, WEB click Data, medical records, biology, engineering. (Web click Data, medical records, etc.)
- Applications can ' t program by hand (no manual programming Application)
- E.g, autonomous helicopter, handwriting recognition, most of Natural Language processing (NLP), computer Vision. (Autonomous helicopters, handwriting recognition, natural language processing, computer vision, etc.)
Self-customizing programs (custom program)
- e.g, Amazon, Netfix product recommendations (for example: recommended products by preference or purchase record)
- Understanding human learning (brain, real AI). (Machine learning is how a sensible class learns.)
What is machine learning