斯坦福大学机器学习第一讲Lecture1

1、 IntroductionWelcomeMachine Learning Machine Learning-Grew out of work in AI-New capability for computersExamples:-Database miningLarge datasets from growth of automation/web.E.g., Web click data, medical records, biology, engineering-Applications cant program by hand.E.g., Autonomous helicopter, ha

2、ndwriting recognition, most of Natural Language Processing (NLP, Computer Vision.-New capability for computersExamples:-Database miningLarge datasets from growth of automation/web.E.g., Web click data, medical records, biology, engineering-Applications cant program by hand.E.g., Autonomous helicopte

3、r, handwriting recognition, most of Natural Language Processing (NLP, Computer Vision.-New capability for computersExamples:-Database miningLarge datasets from growth of automation/web.E.g., Web click data, medical records, biology, engineering-Applications cant program by hand.E.g., Autonomous heli

4、copter, handwriting recognition, most of Natural Language Processing (NLP, Computer Vision.-New capability for computersExamples:-Database miningLarge datasets from growth of automation/web.E.g., Web click data, medical records, biology, engineering-Applications cant program by hand.E.g., Autonomous

5、 helicopter, handwriting recognition, most of Natural Language Processing (NLP, Computer Vision. -Self-customizing programsE.g., Amazon, Netflix product recommendations-New capability for computersExamples:-Database miningLarge datasets from growth of automation/web.E.g., Web click data, medical rec

6、ords, biology, engineering-Applications cant program by hand.E.g., Autonomous helicopter, handwriting recognition, most of Natural Language Processing (NLP, Computer Vision. -Self-customizing programsE.g., Amazon, Netflix product recommendations IntroductionWhat is machinelearningMachine LearningArt

7、hur Samuel (1959. Machine Learning: Field of study that gives computers the ability to learn without being explicitly programmed.Arthur Samuel (1959. Machine Learning: Field of study that gives computers the ability to learn without being explicitly programmed. Machine Learning definitionArthur Samu

8、el (1959. Machine Learning: Field of study that gives computers the ability to learn without being explicitly programmed.Tom Mitchell (1998 Well-posed Learning Problem: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if itsperformance

9、 on T, as measured by P, improves with experience E.Machine Learning definition Classifying emails as spam or not spam. The number (or fraction of emails correctly classified as spam/not spam. None of the abovethis is not a machine learning problem.Suppose your email program watches which emails you

10、 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 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.

11、” Classifying emails as spam or not spam. 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 computer program is said to learn from experience E with respect to some task T and some

12、performance measure P, if its performance on T, as measured by P, improves with experience E.” Classifying emails as spam or not spam. The number (or fraction of emails correctly classified as spam/not spam. None of the abovethis is not a machine learning problem.Suppose your email program watches w

13、hich 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 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 wit

14、h experience E.”Machine learning algorithms:-Supervised learning-Unsupervised learningAlso talk about: Practical advice for applying learning algorithms. IntroductionSupervisedLearningMachine Learning1002003004000500 Housing price prediction.Price ($ in 1000sSize in feet2Predict continuousBreast can

15、cer (malignant, benign Discrete valued output (0 or 1Malignant? 1(Y0(NAge -Clump Thickness-Uniformity of Cell Shape Treat both as classification problems.Treat problem 1 as a classification problem, problem 2 as a regression problem. Treat problem 1 as a regression problem, problem 2 as a classifica

16、tion problem. Treat both as regression problems. Youre running a company, and you want to develop learning algorithms to address each of two problems.Problem 1: You have a large inventory of identical items. You want to predict how many of these items will sell over the next 3 months.Problem 2: Youd

17、 like software to examine individual customer accounts, and for each account decide if it has been hacked/compromised.Andrew Ng Andrew NgIntroduction UnsupervisedLearning Machine LearningAndrew Ngx 1x 2Supervised LearningAndrew NgUnsupervised Learningx 1x 2 Andrew Ng G e n e s Individuals G e n e s

18、Individuals Organize computing clustersImage credit: NASA/JPL-Caltech/E. Churchwell(Univ. of Wisconsin, MadisonCocktail party problem Microphone #1Microphone #2Speaker #1Speaker #2Microphone #1: Microphone #2: Microphone #1: Microphone #2: Audio clips courtesy of Te-Won Lee. Output #1: Output #2: Output #1: Output #2: Andrew Ng Cocktail party problem algorithm W,s,v = svd(repmat(sum(x.*x,1,size(x,1,1.*x*