| Week (Subject to change) |
Topics |
Homework |
Reading |
Special Events and Due Dates |
Slides & Lecture Notes |
| 1 (1/12+14+16) |
Introduction, class logistics, Review: linear algebra |
|
DL Ch. 1 & Ch. 2 |
|
Slides (1), (2) |
2 (1/X+X+X) |
Stochastic processes & probability, distributions, information theory |
|
DL Ch. 3, TEoSL Ch. 1
| HW #0 due X/XX |
Slides (1), (2), (3) |
3 (1/X+X+X) |
Optimization, foundational principles of ML |
|
DL Ch. 4 & Ch. 5 |
|
Slides (1), (2),
(3), (4) |
4 (X/X+X+X) |
Learning theory, generalization, the ML pipeline, K-NN |
|
DL Ch. 5 |
|
Slides (1), (2), (3) |
5 (X/X+X+X) |
Supervised learning: Linear regression |
|
TEoSL Ch. 2.3, 3.1 |
HW #1.1 due X/XX |
Slides (1), (2), (3) |
6 (X/X+X+X) |
Logistic regression (LR) and (2-class) classification |
|
TEoSL Ch. 4.4 |
|
Slides (1), (2), (3) |
7 (X/X+X+X) |
Discriminative modeling with linear classifiers |
|
|
HW #1.2 due X/XX |
Slides (1), (2), (3), (4) |
8 (X/X+X+X) |
Unsupervised learning: dimensionality reduction PCA (Guest lec) |
|
|
|
Slides (1), (2) |
9 (X/X+X+X) |
Probabilistic graphical models (PGMs): naïve Bayes (NB) |
|
NB vs. LR (Ng & Jordan '01) |
HW #2 due X/X |
Slides (1), (2), (3) |
12 (X/X+X, X/X) |
Unsupervised learning, generative models, clustering |
|
| |
Slides (1), (2), (3) |
11 (X/X+X+X) |
Mixtures of Gaussians, decision trees |
|
Random Forests (Breiman '01) |
|
Slides (1), (2), (3) |
13 (X/X+X+X) |
Trees/ensembles, artificial neural networks (ANNs) |
|
DL Ch. 6 |
|
Slides (1), (2), (3) |
12 (X/X+X+X) |
ANNs: reverse-mode differentiation, tricks of the trade |
|
|
|
Slides (1), (2), (3) |
13 (X/X+X+X) |
Violating i.i.d.: RNNs & time-series |
|
DL Ch. 10 |
|
Slides (1) |
14 (X/X+X+X) |
ANNs: Generative modeling |
|
|
Final Exam/Project |
Slides (1), (2) |
16 (X/X, Final: X/XX, XX:XXam-XX:XXpm) |
Outlook, Final Project Presentations |
|
|
| Slides due X/XX, 10:40am |
| (Papers due X/XX, 8am) |
|
|
