# Weekly Schedule

Week-to-week schedule and papers covered are tentative, and may change by the start of the quarter

## Introduction and taxonomy of distribution shifts

- Sep 20
- Introduction
**Lecture**

- Overview of the course
- Distribution shifts in the real world
- A taxonomy of distribution shifts and how they arise

- Sep 22
- Covariate and label shifts
**Lecture**+**Discussion**

- What is a covariate shift?
- Handling covariate shift under distribuitonal overlap.
- Shortcut Learning in Deep Neural Networks

- Sep 27
- Covariate and label shifts 2
**Discussion**

## Domain adaptation theory

- Sep 29
- Domain adaptation
**Lecture**

- When can we provably learn under distribution shift?
- Can unlabeled data help?
- Defining generalization bounds under distribution shift.

- Oct 4
- Domain adaptation 2
**Discussion**

## Neural and representation-based methods

- Oct 6
- Neural domain adaptation
**Lecture**

- Indistinguishability over representations.
- Adversarial approaches to neural domain adaptation.
- Connections to classical theory.

- Oct 11
- Neural domain adaptation 2
**Discussion**

- Oct 13
- Neural domain adaptation 3
**Lecture**

- Provable guarantees from representational indistinguishability
- Self-training based domain adaptation
- Self-supervision based domain adaptation

- Oct 18
- Learning from invariant representations 2
**Discussion**

## Robustness and domain generalization

- Oct 20
- Empirical phenomena in robust machine learning
**Lecture**

- How do different robustness interventions fare in practice?
- Can (data augmentation / unlabeled data / bigger models) help?

- Oct 25
- Empirical phenomena in robust machine learning 2 +
**Project**(**Progress report due**) **Discussion**

- Empirical phenomena in robust machine learning 2 +
- Oct 27
- Connections to causality
**Lecture**

- Distribution shifts as arising from causal interventions.
- Existing connections between causality and robustness.
- Robustness and invariance as tools for causal inference.

- Nov 1
- Connections to causality 2
**Discussion**

- Nov 3
- Minimax methods
**Lecture**

- Robustness as a minimax game between nature and the model.
- Tractable families of worst-case distributions and duality.
- Pitfalls and pessimism from worst-case bounds.

- Nov 8
- Minimax methods 2
**Discussion**

## Adversarial robustness

- Nov 10
- Adversarial examples
**Lecture**

- Defining and motivating adversarial examples.
- Heuristic defenses and their pitfalls
- Provable defenses.

- Nov 15
- Adversarial examples 2
**Discussion**

- Nov 17
- Data poisoning
**Lecture**

- Classical robust statistics
- High-dimensional mean estimation
- Convex optimization under data poisoning

- Nov 29
- Data Poisoning 2
**Discussion**

- Dec 1
- Short project presentations
**Project**