COGS 118A: Natural Computation I (Winter 2012)
MWF 11a1150a at York 3050aSection: fri 1150p HSS1106A
Course website: http://thiscourse.com/ucsd/cogs118a/wi12/
Instructors
Dario Gutierrez (TA)
Office Hours: Th 330430 in CSB115 
Course Description
This is an introductory course in Bayesian statistical methods for modeling natural computation. It will cover topics such as density estimation, linear and logistic regression, Bayesian networks, hidden Markov models, & reinforcement learning. Students will have the opportunity to implement these algorithms in programming exercises, as well as derive some of their properties via mathematical proofs. In addition, the course will touch upon applications of these algorithms in cognitive modeling and data analysis.
Prereqs: Cogs 109, Math 20E,F and Math 180A, or equivalent exposure to programming, linear algebra, calculus and introductory probability theory.
Books
Required

Additional Readings
Probability Theory (E. T. Jaynes, available online)
Information Theory Inference, and Learning (David J.C. Mackay, available online)
Course Details and Policies
Office hours: Wednesdays 24p CSB 157
Programming labs: You may use CSB115, click here for access codes.
Grading: Assignments 35%: late penalty 10%/day. Exams (30%). Final project (35%).
UCSD policy on integrity of scholarship: You must write your own code and homework assignments. You may discuss with other students.
Course Schedule
Week  Date  Topic  Readings  Assignments  Notes 
1  Mon Jan 09  Introduction (slides)  
Wed Jan 11  Probability review  Ch 1.21.24, Linear Algebra Review  
Fri Jan 13  Density estimation  Ch 2.12.2  
2  Mon Jan 16  No class (MLK Jr Day)  Assignment 1  (Assignment 1 due 01/23)  
Wed Jan 18  Gaussian Distribution  Ch 2.3.42.3.6  
Fri Jan 20  Cognitive science applications (slides)  Kording & Wolpert, 2004  
3  Mon Jan 23  Linear regression  Ch 33.1.4  Assignment 2  (Assignment 2 due 01/30) 
Wed Jan 25  Linear regression (slides)  
Fri Jan 27  Biasvariance, Bayesian linear regression  Ch 3.2  3.3  
4  Mon Jan 30  Talk by Andrew Ng  
Wed Feb 01  Bias/variance, Bayesian regression  Ch 3.23.4  Assignment 3 (data) (code)  (Assignment 3 due 02/08) (A1 solutions)  
Fri Feb 03  Classification, Fisher's linear discriminant  Ch 4.14.14  
5  Mon Feb 06  Fisher's discriminant, Probabilistic generative models  Ch 4.1.4, 4.2  
Wed Feb 08  Logistic regression, IRLS, Bayesian Logistic regression  Ch 4.3, 4.5  
Fri Feb 10  Bayesian networks, conditional independence  Ch 8.1, 8.2  Assignment 4 (data) (code)  (Assignment 4 due 02/22)  
6  Mon Feb 13  Bayesian network examples  Project suggestions  
Wed Feb 15  Bayesian Networks: learning & inference  (slides by davies & moore)  
Fri Feb 17  Cognitive modeling: perceptual decisionmaking, conflict resolution (slides)  Gold & Shadlen 2007, Yu et al., 2009  
7  Mon Feb 20  No class (Presidents' day)  
Wed Feb 22  Linear Gaussian models, linear dynamical systems (kalman filter)  (tutorial by Welch & Bishop)  
Fri Feb 24  Kalman filter, cognitive modeling: timescales in motor learning and adaptation (slides)  Kording et al., 2007  A2 solutions A3 solutions 

8  Mon Feb 27  Hidden Markov models: inference  (tutorial by L. Rabiner)  
Wed Feb 29  Hidden Markov models: inference and learning  Assignment 5  (Assignment 5 due mar 7)  
Fri Mar 02  Markov decision processes  (Survey by Kaelbling et al) (additional slides)  
9  Mon Mar 05  MDPs, reinforcement learning  RL simulator  
Wed Mar 07  RL and the brain (slides)  Glaescher et al., 2010, Daw et al., 2005  
Fri Mar 09  Rational models for inhibitory control (slides)  
10  Mon Mar 12  Exam review  
Wed Mar 14  
Fri Mar 16  Project presentations (Regular class time)  Sample Exam Questions  
Fri Mar 16  Project presentations (sections)  Sample exam solutions 