1-3 What is Machine Learning

What is machine learning? In this class we will try to define what it is and also try to give you a sense of it when you want to use machine learning. Even among machine learning prtactitioners there isnt a well accepted definition of what is and what isnt machine learning. But well show you a couple of examples of the ways that people have tried to define it. Heres the definition of what is machine learning due to Arthur Samuel. He defined machine learning as the field of study that gives computers the ability without being explicitly programmed. Samuels claim(聲明) to fame was that back in the 1950s, he wrote a checkers playing program. And the amazing thing about this checkers playing program, was that Arthur Samuel himself, wasnt a very good checkers player. But what he did was, he had to program for it to play 10s of 1000s of games against itself. And by watching what sorts of board(布局) positions tended to lead to wins, and what sorts of board positions tended to lead to loses. The checkers playing program learns over time what are good board positions and what are bad board positions. And eventually learn to play checkers better than Arthur Sameul himself was able to. This was a remarkable result. Although Sameul himself turned out not to be a very good checkers player. But because the computer has the patience to play tens of thousands of games itself. No human, has the patience to play that many games. By doing this the computer was able to get so much checkers-playing experience that it eventually became a better checkers player than Arthur Samuel himself. This is somewhat informal definition, and an older one. Heres a slightly more recent definition by Tom Mitchell, whos a friend out of Carnegie Mellon. So Tom defines machine learning by saying that a well posed learning problem is defined as follows. He says, a computer program is said to learn from experience E, with respect to some task T, and some performance measure P, if its performance on T as measured by P improves with experience E. I actually think he came up with this definition just to make it rhyme. For the checkers playing example the experience E, will be the experience of having the program play 10s of 1000s of games against itself. The task T, will be the task of playing checkers. And the performance measure P, will be the probablity that it wins the next game of checkers against some new opponent(對手).

Throughout these videos, besides teacher(Andrew Ng) trying to teach us study, he will occasionally ask you a qusetion to make sure you understand the content. Heres one, on top is a definition of machine learning by Tom Mitchell

" A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P ,improves with experience E. "

Suppose your email program watches which emails you do or do not mark as spam, and based on that learns how to better filter spam. What is the task T in this setting?

A. Classifying emails as spam or not spam.

B. Watching you label emails as spam or not spam.(E)

C. The number of emails correctly classified as spam/not spam.(P)

The answer is A

In this class The teacher hopes to teach us about various different types of learning algorithms. The main two types are what we call supervised learn and unsupervised learning. He will define what these terms mean more in the next couple videos. But it turns out that in supervised learning, the idea is that were going to teach the computer how to do something, whereas in unsupervised learning were going let it learn by itself. Dont worry if these two terms dont make sense yet, in the next two class He is going to say exactly what these two types of learning are. You will also hear other buzz (嗡嗡聲; 電話) terms such as reinforcement learning(強化學習) and recommender system(推薦系統). These are other types of machine learning algorithms that well talk about later but the two most used types of learning algorithms are probably supervised learning and unsupervised learning and hell define them in the next two class and well spend most of this class talking about these two types of learning algorithms. It turns out one of the other things well spend a lot time on this class is practical advice for applying learning algorithms. This is something that the teacher feels pretty strongly about, and its actually something that he dont know of any other university teaches. Teaching about learning algorithms is like give you a set of tools, and equally important or more important to giving you the tools is to teach you how to apply these tools. He likes to make an analogy to learning to became a carpenter. Image that someone is teaching you how to be a carpenter and they say heres a hammer, heres a screwdriver(螺絲刀), heres a saw(鋸子), good luck. Well, thats no good, right? You, you, you have all these tools, but the more important thing, is to learn how to use tools properly. Theres a huge difference between people that know how to use these machine learning algorithms, versus(與…相對) people who dont know how to use these tools well. Here in Silicon Valley where the teacher(Andrew Ng) lives, when he goes visit different companies very often he sees people are trying to apply machine learning algorithms to some problem and sometimes they have been going at it for six months. But sometimes when he looks at what theyre doing, he says, you know, he could have tould them like, gee, He could have told you six months ago that you should be taking a learning algorithm and applying it in like the slightly modified way and your chance of success would have improved a lot!