I'm a PhD student at Stanford NLP, where I'm rotating with Dorsa Sadigh. I was previously a BMO National Scholar and John H. Moss Scholar at the University of Toronto, where I worked with Graeme Hirst and David Duvenaud.
My research interests are in representation learning for NLP (e.g., word embeddings, additive compositionality) and ML fairness. While at Google, I also worked on graph representation learning and information extraction.
Updates
- (11/17/2019) EMNLP slides for "How Contextual are Contextualized Word Representations?" are now up.
- (09/28/2019) Two papers accepted to EMNLP 2019! Preprints forthcoming.
Publications
-
How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings.
Kawin Ethayarajh.
EMNLP 2019 (talk).
paper code slides bib -
Rotate King to get Queen: Word Relationships as Orthogonal Transformations in Embedding Space.
Kawin Ethayarajh.
EMNLP 2019 (poster).
paper poster bib -
Understanding Undesirable Word Embedding Associations.
Kawin Ethayarajh, David Duvenaud, and Graeme Hirst.
ACL 2019 (talk).
paper blog slides bib -
Towards Understanding Linear Word Analogies.
Kawin Ethayarajh, David Duvenaud, and Graeme Hirst.
ACL 2019 (poster).
paper addendum blog bib -
Unsupervised Random Walk Sentence Embeddings: A Strong but Simple Baseline.
Kawin Ethayarajh.
ACL 2018 - Repl4NLP (talk; best paper).
paper code poster slides screencast bib -
A Rose by Any Other Name: Understanding Judicial Decisions that Do Not Cite Precedent.
Kawin Ethayarajh, Andrew Green, and Albert Yoon.
Journal of Empirical Legal Studies. 15(3): 563-596.
paper -
The Effect of Photoperiod on the Mood of Reddit Users.
Kawin Ethayarajh and Frank Rudzicz.
Cyberpsychology, Behavior, and Social Networking. 20(4): 238-245.
paper press