Aditya Yedetore

email: [last]@bu.edu

I am a graduate student in the Linguistics Department at Boston University, working with Professor Najoung Kim. I am interested in semantics, syntax, and language acquisition. My current work involves comparing linguistic generalization in neural networks and children.

Previously I completed a double major in Cognitive Science and Computer Science at Johns Hopkins University. I did research with Tal Linzen, Bob Frank, and Tom McCoy studying language acquisition through simulation with deep learning models. I also did research with Justin Halberda and Nico Arlotti on the interface between our visual system and the meanings of “All” and “Each”.

news

Oct 8, 2024	New preprint! With Najoung Kim.
Sep 24, 2024	Presented at Harvard’s Language and Cognition talk series.
May 2, 2024	Invited talk at the MIT CPL on cues to hierarchical generalization in neural networks.
Apr 16, 2024	Spotlight talk & poster at NENLP 2024 on semantic bootstrapping in neural networks.
Jul 10, 2023	Presented at ACL 2023 on hierarchical generalization in neural networks.

selected publications

ACL

How poor is the stimulus? Evaluating hierarchical generalization in neural networks trained on child-directed speech

Aditya Yedetore, Tal Linzen, Robert Frank, and 1 more author

In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Jul 2023

Abs arXiv PDF Poster

When acquiring syntax, children consistently choose hierarchical rules over competing non-hierarchical possibilities. Is this preference due to a learning bias for hierarchical structure, or due to more general biases that interact with hierarchical cues in children’s linguistic input? We explore these possibilities by training LSTMs and Transformers - two types of neural networks without a hierarchical bias - on data similar in quantity and content to children’s linguistic input: text from the CHILDES corpus. We then evaluate what these models have learned about English yes/no questions, a phenomenon for which hierarchical structure is crucial. We find that, though they perform well at capturing the surface statistics of child-directed speech (as measured by perplexity), both model types generalize in a way more consistent with an incorrect linear rule than the correct hierarchical rule. These results suggest that human-like generalization from text alone requires stronger biases than the general sequence-processing biases of standard neural network architectures.