About

I’m a third-year Ph.D. student in the Department of Linguistics at New York University. I study how neural language models learn language, and how that is similar to or different from how humans acquire language. Before coming to NYU, I earned my bachelors degree in Linguistics from Yale University; my undergraduate thesis on algebraic generalization in neural networks was advised by Bob Frank as part of the CLAY Lab.

Research Interests

Applications of pure mathematics to linguistics and machine learning. How can we precisely describe how difficult a given task is for a neural language model? What are the mathematical properties of natural language when viewed as an algebraic object? Can we quantify how difficult natural language is for a given neural architecture?
Properties of neural network generalization. How do neural networks generalize to new inputs? What inductive biases do neural architectures have, and what biases induce human-like generalization? How can we build neural networks with data efficiency comparable to humans? Can we guarantee that neural networks will generalize in a consistent, interpretable, and safe manner? How do neural networks learn to generalize compositionally?
NLP for Jewish languages. How can we build the best NLP systems for Hebrew, Yiddish, and Aramaic? How can we improve tokenization for non-concatenative morphology?