Probing the Representational Structure of Regular Polysemy via Sense Analogy Questions: Insights from Contextual Word Vectors
Cognitive Science · 2024-03 · 48 (3)
DOI: 10.1111/cogs.13416 ↗At a glance
Journal article in Cognitive Science, published online on 2024-03-14. Written by Jiangtian Li and 1 co-authors. Institution: Department of Psychology, University of Toronto. Major MEDLINE topics: Psycholinguistics, Semantics. Cited 5 times so far (Crossref); 96 references. The full text is freely available (open access, CC BY 4.0). Funded by: Natural Sciences and Engineering Research Council of Canada.
Abstract (original language)
Regular polysemes are sets of ambiguous words that all share the same relationship between their meanings, such as CHICKEN and LOBSTER both referring to an animal or its meat. To probe how a distributional semantic model, here exemplified by bidirectional encoder representations from transformers (BERT), represents regular polysemy, we analyzed whether its embeddings support answering sense analogy questions similar to “is the mapping between CHICKEN (as an animal) and CHICKEN (as a meat) similar to that which maps between LOBSTER (as an animal) to LOBSTER (as a meat)?” We did so using the LRcos model, which combines a logistic regression classifier of different categories (e.g., animal vs. meat) with a measure of cosine similarity. We found that (a) the model was sensitive to the shared structure within a given regular relationship; (b) the shared structure varies across different regular relationships (e.g., animal/meat vs. location/organization), potentially reflective of a “regularity continuum;” (c) some high‐order latent structure is shared across different regular relationships, suggestive of a similar latent structure across different types of relationships; and (d) there is a lack of evidence for the aforementioned effects being explained by meaning overlap. Lastly, we found that both components of the LRcos model made important contributions to accurate responding and that a variation of this method could yield an accuracy boost of 10% in answering sense analogy questions. These findings enrich previous theoretical work on regular polysemy with a computationally explicit theory and methods, and provide evidence for an important organizational principle for the mental lexicon and the broader conceptual knowledge system.
Subjects
MeSH terms (MEDLINE)
Author keywords
Authors and institutions
- Department of Psychology, University of Toronto.
- Department of Psychology and Department of Computer Science, University of Toronto BCBL, Basque Center on Cognition, Brain, and Language.
ORCID profiles: Jiangtian Li
Funding
- Natural Sciences and Engineering Research Council of Canada
Cite
APA
Li, J., & Armstrong, B. C. (2024). Probing the Representational Structure of Regular Polysemy via Sense Analogy Questions: Insights from Contextual Word Vectors. Cognitive Science, 48(3). https://doi.org/10.1111/cogs.13416
Vancouver
Li J, Armstrong BC. Probing the Representational Structure of Regular Polysemy via Sense Analogy Questions: Insights from Contextual Word Vectors. Cognitive Science. 2024;48(3). doi:10.1111/cogs.13416
BibTeX
@article{Li2024,
author = {Li, Jiangtian and Armstrong, Blair C},
title = {Probing the Representational Structure of Regular Polysemy via Sense Analogy Questions: Insights from Contextual Word Vectors},
journal = {Cognitive Science},
year = {2024},
volume = {48},
number = {3},
doi = {10.1111/cogs.13416}
}
EUVIMED