1/5 |
Introductions, Logistics |
|
1/10 |
Introductory lecture |
|
1/12 |
Sparse, Dense, and Attentional Representations for Text Retrieval |
Bhuwan & Sam |
1/17 |
MLK Jr. Day; no class |
|
1/19 |
Nearest Neighbor Machine Translation |
Group 1 |
1/24 |
Learning with Instance Bundles for Reading Comprehension |
Group 2 |
1/26 |
Facts as Experts: Adaptable and Interpretable Neural Memory over Symbolic Knowledge |
Group 3 |
1/31 |
FNet: Mixing Tokens with Fourier Transforms |
Group 1 |
2/2 |
When Attention Meets Fast Recurrence: Training Language Models … |
Group 2 |
2/7 |
DEMix Layers: Disentangling Domains for Modular Language Modeling |
Group 3 |
2/9 |
Bad Characters: Imperceptible NLP Attacks |
Group 1 |
2/14 |
Distributionally Robust Language Modeling |
Group 2 |
2/16 |
Counterfactual Invariance to Spurious Correlations in Text Classification |
Group 3 |
2/21 |
Achieving Model Robustness through Discrete Adversarial Training |
Group 1 |
2/23 |
Learning to Recombine and Resample Data for Compositional … |
Group 2 |
2/28 |
Active Learning by Acquiring Contrastive Examples |
Group 3 |
3/2 |
Neural Data Augmentation via Example Extrapolation |
Group 1 |
3/7 & 3/9 |
Spring break; no class |
|
3/14 |
Counterfactual Data Augmentation for Neural Machine Translation |
Group 2 |
3/16 |
Learning to Faithfully Rationalize by Construction |
Group 3 |
3/21 |
Measuring Association Between Labels and Free-Text Rationales |
Group 1 |
3/23 |
Aligning Faithful Interpretations with their Social Attribution |
Group 2 |
3/28 |
FastIF: Scalable Influence Functions for Efficient Model |
Group 3 |
3/30 |
QED: A Framework and Dataset for Explanations in Question Answering |
Group 1 |
4/4 |
Memorizing Transformers |
Group 2 |
4/6 |
LM-Critic: Language Models for Unsupervised Grammatical Error Correction |
Group 3 |
4/11 |
Presentations |
TBD |
4/13 |
Presentations |
TBD |