Hsuan-Tien Lin > Courses > Machine Learning, Fall 2008

Hsuan-Tien Lin

Home | MOOCs | AIsk | Courses | Research Group | Awards | Publications | Presentations | Programs/Data

Machine Learning, Fall 2008

Course Description

Machine learning allows computational systems to adaptively improve their performance with experience accumulated from the data observed. This course introduces the basics of learning theories, the design and analysis of learning algorithms, and some applications of machine learning.

Course Information

Instructor: Hsuan-Tien Lin (htlin AT csie DOT ntu DOT edu DOT tw)
office hour: after class, R105, or by appointment, R314
TA: Hanhsing Tu and Ming-Feng Tsai (ml08fall AT csie DOT ntu DOT edu DOT tw)
TA hour: 12:00--14:00 on the day before homework due, R106
Time: Thursdays 14:20--17:20 (we will take a long break around 15:50)
Room: R105 of the CSIE building
Textbook (for reference only, not required): The Elements of Statistical Learning, by T. Hastie, R. Tibshirani, J. H. Friedman (2001)
Language: English
Grading: 70% homework, 30% project (tentative)
Discussion Forum: provided by CEIBA here. You need to login from CEIBA to make a post.

Announcements

homework 0 released on 09/15/2008, no due date (self-graded); see the email annoucement from the instructor
class policy released on 09/18/2008; please read it very carefully and check the in-class slides
if you want to be manually added to the class, please send the instructor an email illustrating why you should be the "special case" (see explanations here)
a special TA session for homework 0 will be held on 09/25/2008, 12:00--14:00, R106; see the email announcement from the instructor
a bug in homework 0 has been fixed; thanks to Mong-Jen for pointing it out; please check the discussion forum for details.
homework 1 released on 09/25/2008, to be due on 10/02/2008; some guidelines about submission and source code uploading also announced
a brief description of the PLR algorithm is put online to help everyone conquer homework 1; see the email announcement from the instructor
The due date of homework 1 is extended to 10/06/2008, 4pm, in Room 526.
homework 2 released on 10/09/2008, to be due on 10/16/2008
homework 3 released on 10/23/2008, to be due on 10/30/2008
homework 4 released on 11/06/2008, to be due on 11/20/2008
homework 5 released on 11/27/2008, to be due on 12/04/2008
some corrections of homework 5 announced on 12/01/2008 and 12/03/2008
homework 6 released on 12/11/2008, to be due on 12/18/2008
some corrections of homework 6 announced on 12/13/2008 and 12/15/2008
final project released on 12/18/2008, to be due on 1/15/2009
homework 7 released on 12/25/2008, to be due on 1/8/2009
the solution of 5.2(3) posted on 12/27/2008
some corrections of homework 7 announced on 1/5/2009
the evaluation result is here (uploaded 5/17/2009). Thank you all for the valuable suggestions.
Congratulations to our two TAs for winning the excellent TA award in the CSIE department!! They are indeed EXCELLENT.

Class Policy

main policy: see here
homework due at 16:00 in class (see schedule below)
you are allowed to turn in some of the problems separately (but please mark your name clearly!), and only the late parts will be penalized
you must write your homework in English

Course Plan

date	syllabus	todo/done	suggested reading
2008.09.18	introduction/unlearnability	homework 0 released; policy explained and released	textbook, Ch 1, Sec 2.1, Sec 2.2 a brief introduction to machine learning
2008.09.25	learnability/perceptrons	homework 1 released; homework guidelines announced	textbook, Subsec 2.6.1, Subsec 2.6.2 The PLR algorithm handout Sec 1 and Sec 2 of Abu-Mostafa, 1989 (note the slightly different definitions of pi and nu; other Secs would be taught later)
2008.10.02	perceptron/linear regression	homework 1 due	textbook, Subsec 2.3.1, Sec 3.2
2008.10.09	optimization/generalization	homework 2 released	textbook, Sec 4.4, Subsec 4.5.1, page 196 gradient descent and stochastic gradient descent
2008.10.16	generalization/VC inequality	homework 2 due	textbook, Sec 7.9 proof of the Sauer's lemma (i.e., bounds on B(N, M)); see here for a similar proof proof of the VC inequality Sec 3 of Abu-Mostafa, 1989 (note the slightly different definitions of pi and nu)
2008.10.23	generalization/nearest neighbors	homework 3 released	textbook, Subsec 2.3.2, Sec 7.10, Sec 13.3 proof of the Sauer's lemma (i.e., bounds on B(N, M)); see here for a similar proof proof of the VC inequality
2008.10.30	nearest neighbors/RBF Network/generalized linear model	homework 3 due	textbook, Sec 5.1, Sec 6.7, Sec 13.3, Subsec 14.3.6
2008.11.06	Neural Network	homework 4 released	textbook, Sec 11.3, Sec 11.4 backpropagation algorithm
2008.11.13	Neural Network/Bayesian	(midterm week)	textbook, Sec 11.3, Sec 11.4
2008.11.20	Bayesian	homework 4 due	textbook, Sec 8.3
2008.11.27	Bayesian/Ensemble	homework 5 released	textbook, Sec 8.3
2008.12.04	Bagging/Boosting	homework 5 due	textbook, Sec 8.7, Ch 10 slides on AdaBoost
2008.12.11	Boosting	homework 6 released	textbook, Ch 10 slides on AdaBoost A short introduction to boosting
2008.12.18	SVM	homework 6 due; final project announced	textbook, Sec 12.1, 12.2, 12.3
2008.12.25	SVM	homework 7 released	textbook, Sec 12.1, 12.2, 12.3 some notes for deriving dual SVM (similar to what we did in class) a short SVM tutorial a practical guide to support vector classification
2009.01.01	no class	New Year's Day (holiday)
2009.01.08	summary; Q and A	homework 7 due
2009.01.15	no class	(final week); final project due

Last updated at CST 13:07, October 04, 2023
Please feel free to contact me: htlin.email.png