Compositional and lexical semantics in RoBERTa, BERT and DistilBERT: A case study on CoQA

17Citations
Citations of this article
131Readers
Mendeley users who have this article in their library.

Abstract

Many NLP tasks have benefited from transferring knowledge from contextualized word embeddings, however the picture of what type of knowledge is transferred is incomplete. This paper studies the types of linguistic phenomena accounted for by language models in the context of a Conversational Question Answering (CoQA) task. We identify the problematic areas for the finetuned RoBERTa, BERT and DistilBERT models through systematic error analysis - basic arithmetic (counting phrases), compositional semantics (negation and Semantic Role Labeling), and lexical semantics (surprisal and antonymy). When enhanced with the relevant linguistic knowledge through multitask learning, the models improve in performance. Ensembles of the enhanced models yield a boost between 2.2 and 2.7 points in F1 score overall, and up to 42.1 points in F1 on the hardest question classes. The results show differences in ability to represent compositional and lexical information between RoBERTa, BERT and DistilBERT.

Cite

CITATION STYLE

APA

Staliūnaité, I., & Iacobacci, I. (2020). Compositional and lexical semantics in RoBERTa, BERT and DistilBERT: A case study on CoQA. In EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference (pp. 7046–7056). Association for Computational Linguistics (ACL). https://doi.org/10.18653/v1/2020.emnlp-main.573

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free