Comparing Text Representations: A Theory-Driven Approach

Gregory Yauney, David Mimno


Abstract
Much of the progress in contemporary NLP has come from learning representations, such as masked language model (MLM) contextual embeddings, that turn challenging problems into simple classification tasks. But how do we quantify and explain this effect? We adapt general tools from computational learning theory to fit the specific characteristics of text datasets and present a method to evaluate the compatibility between representations and tasks. Even though many tasks can be easily solved with simple bag-of-words (BOW) representations, BOW does poorly on hard natural language inference tasks. For one such task we find that BOW cannot distinguish between real and randomized labelings, while pre-trained MLM representations show 72x greater distinction between real and random labelings than BOW. This method provides a calibrated, quantitative measure of the difficulty of a classification-based NLP task, enabling comparisons between representations without requiring empirical evaluations that may be sensitive to initializations and hyperparameters. The method provides a fresh perspective on the patterns in a dataset and the alignment of those patterns with specific labels.
Anthology ID:
2021.emnlp-main.449
Volume:
Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2021
Address:
Online and Punta Cana, Dominican Republic
Editors:
Marie-Francine Moens, Xuanjing Huang, Lucia Specia, Scott Wen-tau Yih
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
5527–5539
Language:
URL:
https://aclanthology.org/2021.emnlp-main.449
DOI:
10.18653/v1/2021.emnlp-main.449
Bibkey:
Cite (ACL):
Gregory Yauney and David Mimno. 2021. Comparing Text Representations: A Theory-Driven Approach. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 5527–5539, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
Comparing Text Representations: A Theory-Driven Approach (Yauney & Mimno, EMNLP 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.emnlp-main.449.pdf
Video:
 https://aclanthology.org/2021.emnlp-main.449.mp4
Code
 gyauney/data-label-alignment
Data
GLUEMNISTMultiNLIQNLISNLI