Modeling and Data Analysis - COGS 109 Fall 2010

Warren Lecture Hall 2005 : M/W/F 11-11:50am

Note see wiki for review session info as well as some sample final questions/solutions and an text page on why hidden units help (and info of HW policy)

See note on Nov 29 entry about HW5

All questions should be posted to the

course wiki.

Students with one-time conflicts (e.g. sickness, holidays) may attend other sections as needed. Please try to target the less full sections.

Students in need of extra help, may attend extra sections as needed (space permitting)

Cheating will not be tolerated: University Academic Dishonesty Policy

Current Grades (including Midterm) will be available on webct

Midterm1 Stats from 2008 Midterm2 Stats from 2008

Professor: Virginia de Sa (SSRB/CRB (not CSB) 2nd floor phone 2-5095)

TAs: Ben Cipollini, Crane Huang

Office hours/locations will be listed on the course wiki

Recommended Text: Matlab for Engineers Explained (Gustafsson and Bergman) -- Amazon may have for $32.54)

Grading Scheme: Midterms, Final Exam, approx 5 programming assignments (30%). We will downweight the midterms if you do better on the final. IF you hand in all assignments, we will remove the lowest grade. But you must hand in all assignments for this privilege.

For access code to CSB 115

For your account login info
Section info to go here (All Sections in CSB 115)

Final exam will be Tuesday December 7th

Useful References

Online Matlab Tutorials

Course Goals

This is a relatively new course designed to be both an
A) ``end of the road computation course'' for non-Computational majors
B) an introduction to serious Computational Methods for those taking the 118(A/B) courses.

Doing both well involves a somewhat difficult compromise but there are several aspects to both
-Improve your programming skills
-Introduce you to the wonders of Matlab
-Give you a flavor for Cognitive Science data analysis and modeling applications (many of which will explained in mathematical depth in 118)

Tentative Course Schedule Fall 2010B
Date	Lecture Topic	Slides	Reading	Week's Section Topic	Homework Notes
Fri Sept 24	Introduction to COGS 109	Slides here		NO Sections	OPEN MATLAB and type demo and explore some of the getting started videos
Mon Sept 27	Intro to Linear Algebra	Notes here	free online linear algebra notes	Math Review/Help
Weds Sept 29	Intro to Matlab	Notes here	Preface, pp1-8, 12-15, 45-52 (Matlab book), Look over appendices Derivatives (helpful for section and homework) , nice online tutorial , Look over what is available under the Matlab tutorials link above	Math Review/Help
Fri Oct 1	Matlab Functions and graphing	Notes here	pp 41-56, pp 22-26, Appendix E		Homework 1 functions and derivatives HW1 Matrix and Vector Algebra
Mon Oct 4	Matlab Functions and graphing	see notes above	pp 41-56, pp 22-26, Appendix E
Wed Oct 6	Probability refresher, Bayes rule	Notes here
Fri Oct 8	Using the Matlab programming environment - Josh Lewis
Mon Oct 11	Data Visualization	notes here	There is a fair amount of online reading embedded in the notes above
Wed Oct 13	BCI's Neurosky guest speaker
Fri Oct 15	Clustering, Kmeans	Slides here			HW2 NOW POSTED
Mon Oct 18	stats issues	hypothesis testing comparetests.m	pp 127-131 (text)
Wed Oct 20	filtering	filtering filterdemo.m
Fri Oct 22	filtering cont'd
Mon Oct 25	PCA	HW2 DUE at beginning of class -- Bring your HW to class PCA slides now here supplemental covariance slides (called from above slides
Wed Oct 27	cont'd from above	pca2dexample.m faceexample09.m viewcolumn.m
Fri Oct 29	REVIEW FOR MIDTERM	midterm 1 supplemental study notes
Mon Nov 1	Midterm 1
Wed Nov 3	PCA cont'd	faceexample09.m hw3bdata.mat pcaexample2d.m viewcolumn.m eigsort.m
Fri Nov 5	Cont'd from above				HW3 now available -- see advice notes and helper functions (eigsort.m, viewcolumn.m) on wiki (helper functions also above)
Mon Nov 8	Linear Regression, overfitting, non-linear function fitting	Notes now here predyval.m (used in the notes) linregress.m (extracted code from notes) linregresslargerexample.m (extracted code from notes) leaveoutcode.m (extracted code from notes) o
Wed Nov 10	Nonlinear function fitting	Updated notes now here myfitdemoslow.m myfitfunslow.m myfitdemo.m myfitfun.m
Fri Nov 12	Introduction to Neural Networks	Notes here (minus graphs drawn in class)
Mon Nov 15	Gradient Descent	Notes now here (minus graphs and derivations in class)			HW4 now available
Wed Nov 17	Multi-layer perceptrons	Notes to be available (minus graphs and derivations in class) neuralnetfit.m neuralfitfun.m neuralnetfit2.m neuralfitfun2.m
Fri Nov 19	Training Networks in Matlab	Notes now here xorbplay.m xorbplayver71.m	Backpropagation section of the Neural Network Toolbox hDocumentation (available from the help system under Neural Network Toolbox) or online at http://www.mathworks.com/access/helpdesk/help/toolbox/nnet/backprop.html#backpropagation READ NOW Concentrate on Intro, Fundamentals
Mon Nov 22	Training issues in Neural Networks	Notes now here
Wed Nov 24	Review for Midterm 2	Sample midterm 2 questions here (updated 11/27 with PCA questions at top) nnet.jpg (used in sample midterm) nnet4.jpg (used in sample midterm) Sample MT2 solutions now posted (11/27) NOTE REVIEW NOTES FOR WHOLE COURSE ARE BELOW -- FOR MIDTERM 2 KNOW PCA through PERCEPTRONS (see delineation in the review notes)
Mon Nov 29	Neural Networks and Review for MT2 cont'd				HW5 (Do not panic -- this should be considered a study aid for the final exam (not MT2) - solutions will be provided before it is due at the final exam)
Wed Dec 1	Midterm 2
Fri Dec 3	Last Class - Review	review notes now here