We find that lexica induced from context-oblivious models are more generalizable and interpretable than those from more accurate context-aware transformer models. In addition, lexicon scores can identify explanatory words more reliably than a high performing transformer with feature-importance measures like SHAP.
To compare the degree of generalizability and interpretability of the lexica induced from context- aware or context-oblivious models and to gain in- sights into ...
This work introduces a new research problem for the NLP community: how to measure variation in cultural constructs across regions using language and ...
Our method builds upon a line of work in NLP called lexicon induction (Araque et al., 2020;Buechel et al., 2020; Geng et al., 2022; Havaldar et al., 2023a), ...
Inducing Generalizable and Interpretable Lexica. Download data. wget http ... We identify generalizability and interpretability as two essential properties of ...
Inducing generalizable and interpretable lexica. Y Geng, Z Wu, R Santhosh, T Srivastava, L Ungar, J Sedoc. Findings of the Association for Computational ...
Yilin Geng, Zetian Wu, Roshan Santhosh, Tejas Srivastava, Lyle H. Ungar, João Sedoc: Inducing Generalizable and Interpretable Lexica.
Co-authors ; Inducing generalizable and interpretable lexica. Y Geng, Z Wu, R Santhosh, T Srivastava, L Ungar, J Sedoc. Findings of the Association for ...
Jun 17, 2024 · We build knowledge-guided lexica to model cultural variation. Our method encodes domain knowledge via seed words based on cultural psychology theory.
Inducing Generalizable and Interpretable Lexica. Yilin Geng, Zetian Wu, Roshan Santhosh, Tejas Srivastava, Lyle Ungar, João Sedoc Findings of the Association ...