The following is a preliminary schedule of the lectures and presentations for CS 4033/5033 Machine Learning for Fall 2009. This specific order of each topic will depend on the class projects and interest. This will be finalized in the first week. The schedule will be updated throughout the semester as we delve deeper into special topics and/or add additional topics of interest. The project deadlines are in red and will not change, even if other parts of the schedule change. Major new topics are noted in blue.
| Date | Topic | Assigned Reading | Assigned | Due |
| Aug 25 (Week 1) | What is machine learning? What will I learn if I take this class? Types of machine learning. | |||
| Aug 27 Reinforcement Learning | Introduction to RL, Project introduction | Chapter 1 (Sutton & Barto) | Project | Pretest |
| Sep 1 (Week 2) | Oral project proposals |
HW 1 | ||
| Sep 3 | Exploration & Exploitation | 2.1-2.3, 2.5-2.7 (Sutton & Barto) | ||
| Sep 8 (Week 3) | The RL problem | Chapter 3 (Sutton & Barto) |
HW 2 | HW 1, Written proposal |
| Sep 10 | The RL problem | |||
| Sep 15 (Week 4) | The RL problem | HW 2 | ||
| Sep 17 | Dynamic Programming | Chapter 4 (Sutton & Barto) | HW 3 | |
| Sep 22 (Week 5) | Temporal Difference learning | 6.1-6.5, 6.8 (Sutton & Barto) |
Checkpoint 1 | |
| Sep 24 | Eligibility Traces | Ch 7 (Sutton & Barto) | HW 3 | |
| Sep 29 (Week 6) Supervised Learning | Nearest neighbor methods, Regression, Locally weighted regression | Section 20.4 in Russell & Norvig, two wikipedia links in previous box | HW 4 | Checkpoint 2 Meeting: Handwriting, Eclairs, Tornado, Risk, Scheduling Written: Flu, Text, Turbulence, Oral: Mario |
| Oct 1 | Introduction to Neural Networks, , Logistic regression |
Section 20.5 in Russell & Norvig) | ||
| Oct 6 (Week 7) | Training the network, Backprop |
HW 4 | ||
| Oct 8 | Backprop | Checkpoint 3 Meeting: Text, Turbulence, Mario Written: Handwriting, Eclairs, Risk, Scheduling Oral: Flu, Tornado |
||
| Oct 13 (Week 8) | Homework 5 in class group work on backprop and neural nets
|
HW 5 | HW 5 | |
| Oct 15 | Advanced topics in neural nets including function approximation with RL |
RL/NN (see RL Book 11.1), also 8.1-8.4 in RL book | ||
Oct 20 (Week 9) General techniques |
Function approximation with RL | HW 6 | Checkpoint 4 Oral: Scheduling, Turbulence, Risk |
|
| Oct 22 | Overfitting/Model Complexity | pages 660-663 of Russell & Norvig (Chapter 18) | ||
| Oct 27 (Week 10) Graphical Models | Introduction to Bayesian Networks
|
14.1 ( Russell & Norvig) 18.4 (Russell & Norvig) |
HW 6 | |
| Oct 29 | Naive Bayes | Ch 8 (Mitchell sample chapter) | Checkpoint 5 Meeting: Flu, Mario, Written: Tornado, Turbulence, Risk, Scheduling Oral: Eclairs, Text |
|
| Nov 3 (Week 11) | Bayesian Networks |
HW 7 | ||
| Nov 5 | Conditional Independence | 14.2 | ||
| Nov 10 (Week 12) | Exact inference | 14.4 (Russell & Norvig) | Checkpoint 6 Meeting: Eclairs, Neterpillars, Turbulence, Tornado, Text, Risk, Scheduling Written: Flu, Mario |
|
| Nov 12 | Exact inference in-class exercises | HW 8 in class | HW 7, HW 8 | |
| Nov 17 (Week 13) | Hybrid networks | 14.3 | ||
| Nov 19 | Mixture models, Expectation Maximization | Section 20.3 (Russell & Norvig) except for the Bayes Net part |
HW 9 | |
| Nov 24 (Week 14) | EM for Bayesian Networks (learning structure) | Preliminary writeup | ||
| Nov 26 | Thanksgiving vacation |
|||
| Dec 1 (Week 15) | Ensemble methods | Peer reviews | ||
| Dec 3 | ||||
| Dec 8 (Week 16) | Class cancelled |
HW 9 | ||
| Dec 10 | Project presentations? |
Final writeup due | ||
| December 11, 2-5pm | Poster presentation: Sarkeys A & B |
|||
| Dec 17 (Final exam period 8-10am) | Project presentations |
|||