========
| Week: date | Topic(s) | Course | Readings | Additional information | |||
|---|---|---|---|---|---|---|---|
| 1: 1/17 | Intro to ML and Policy | Slides | |||||
| 2: 1/22 | Logistic regression | Slides | SLwS 3, ESL 4.4, MLPP 8 | 701; Greene 18.2 | |||
| 2: 1/24 | Evaluating prediction | Slides | On confusion matrices and ROC curves | ||||
| 3: 1/29 | Discussion 1: academics and policy-makers on ML in Policy | Write-up Question and answer templates; example | Readings; sign up | ||||
| 3: 1/31 | Variants: linear regression, regularization | Slides | ESL 3, SLwS 2, ESL 5 | 701 | |||
| 4: 2/5 | Ensembles, forests, boosting, and gradient boosting | Slides | ESL 10.1, 10.9, 16 | ||||
| 4: 2/7 | Graphical models | HW 1 link; Slides | Bishop 8, Murphy 2, 10 | 701, Traffic flows, Public participation and groundwater contamination | |||
| 5: 2/12 | Missing at random data | Slides | Murphy 11.4, 11.6; Murphy 10 | HW 2 to be released | |||
| 5: 2/14 | Missing not at random data | Murphy 8.6.2 | |||||
| 6: 2/19 | Causality and DAGs | Slides | ICML 2016 tutorial slides, IPTW derivation | ||||
| 6: 2/21 | Causality and reweighting/matching | Slides | |||||
| 7: 2/26 | Machine learning and fairness | HW 2 due | Statistical fairness - Mitchell and Shadlen | ||||
| 7: 2/28 | Discussion 2: ML workshops on Fairness | Discussion 2 sign-up | |||||
| 8: 3/5 | Neural networks | Slides | Murphy 16.5 | HW 3 to be released | |||
| 8: 3/7 | Intro to deep learning | Murphy 28 | |||||
| 9: 3/12 and 3/14 | ---- Spring break: no class ---- | ||||||
| 10: 3/19 | Deep learning | Murphy 28; Deep Learning 2017 (slides) | |||||
| 10: 3/21 | Deep learning | HW 3 due | |||||
| 11: 3/26 | Fitzpatrick, on ML for predictive policing | ||||||
| 11: 3/28 | Survival analysis | CASI 9 | Cox derivation | ||||
| 12: 4/2 | Hawkes processes | Proposals due | Hawkes tutorial | ||||
| 12: 4/4 | Kernels and support vector machines | Murphy 14; CASI 19 | HW 4 to be released | ||||
| 13: 4/9 | Language modeling | Murphy 27.3; word2vec | |||||
| 13: 4/11 | --> --> | Midterm | |||||
| 14: 4/16 | Markov logic networks | Murphy 27.5, 27.6 | |||||
| 14: 4/18 | Discussion 3: Relational learning | ||||||
| Tentative schedule follows | |||||||
| 15: 4/23 | Dimensionality reduction | HW 4 due | ESL 14.5, Manifolds | ||||
| 15: 4/25 | Discussion 4: TBD | ||||||
| 16: 4/30 | Reinforcement learning | ||||||
| 16: 5/2 | Discussion 5: TBD | Project papers due |
Note: bold indicates a deadline
Slides and solutions will be made available on Canvas.