Contextual Embeddings: When Are They Worth It?

Simran Arora, Avner May, Jian Zhang, Christopher Ré


Abstract
We study the settings for which deep contextual embeddings (e.g., BERT) give large improvements in performance relative to classic pretrained embeddings (e.g., GloVe), and an even simpler baseline—random word embeddings—focusing on the impact of the training set size and the linguistic properties of the task. Surprisingly, we find that both of these simpler baselines can match contextual embeddings on industry-scale data, and often perform within 5 to 10% accuracy (absolute) on benchmark tasks. Furthermore, we identify properties of data for which contextual embeddings give particularly large gains: language containing complex structure, ambiguous word usage, and words unseen in training.
Anthology ID:
2020.acl-main.236
Volume:
Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics
Month:
July
Year:
2020
Address:
Online
Editors:
Dan Jurafsky, Joyce Chai, Natalie Schluter, Joel Tetreault
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2650–2663
Language:
URL:
https://aclanthology.org/2020.acl-main.236
DOI:
10.18653/v1/2020.acl-main.236
Bibkey:
Cite (ACL):
Simran Arora, Avner May, Jian Zhang, and Christopher Ré. 2020. Contextual Embeddings: When Are They Worth It?. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 2650–2663, Online. Association for Computational Linguistics.
Cite (Informal):
Contextual Embeddings: When Are They Worth It? (Arora et al., ACL 2020)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingest-2024-clasp/2020.acl-main.236.pdf
Video:
 http://slideslive.com/38929148
Data
CoNLL 2003GLUE