Completed
One assumption is test = train
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Out of Distribution Generalization
Automatically move to the next video in the Classroom when playback concludes
- 1 Classifying cows versus penguins
- 2 A common issue
- 3 Example: cows v. penguins
- 4 Idea 1: nuisance randomization
- 5 Doesn't work by itself
- 6 Idea 2: Uncorrelating representations
- 7 What's the best uncorrelating representation?
- 8 A simulation
- 9 Waterbirds versus landbirds
- 10 Pneumonia classification
- 11 Related work
- 12 What did I mean by local optima?
- 13 Did we lose anything?
- 14 Natural language inference
- 15 Generic setup not solvable
- 16 One assumption is test = train
- 17 Invariant coupling
- 18 Subgroup Coupling
- 19 Common factor coupling
- 20 What does good mean?
- 21 Simultaneous Optimality?
- 22 Why does causality appear?
- 23 The added flexibility comes with an estimation cost
- 24 Some questions
- 25 Research References