Empirical Evaluation of Pre-trained Transformers for Human-Level NLP: The Role of Sample Size and Dimensionality

Adithya V Ganesan, Matthew Matero, Aravind Reddy Ravula, Huy Vu, H. Andrew Schwartz


Abstract
In human-level NLP tasks, such as predicting mental health, personality, or demographics, the number of observations is often smaller than the standard 768+ hidden state sizes of each layer within modern transformer-based language models, limiting the ability to effectively leverage transformers. Here, we provide a systematic study on the role of dimension reduction methods (principal components analysis, factorization techniques, or multi-layer auto-encoders) as well as the dimensionality of embedding vectors and sample sizes as a function of predictive performance. We first find that fine-tuning large models with a limited amount of data pose a significant difficulty which can be overcome with a pre-trained dimension reduction regime. RoBERTa consistently achieves top performance in human-level tasks, with PCA giving benefit over other reduction methods in better handling users that write longer texts. Finally, we observe that a majority of the tasks achieve results comparable to the best performance with just 1/12 of the embedding dimensions.
Anthology ID:
2021.naacl-main.357
Volume:
Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Month:
June
Year:
2021
Address:
Online
Editors:
Kristina Toutanova, Anna Rumshisky, Luke Zettlemoyer, Dilek Hakkani-Tur, Iz Beltagy, Steven Bethard, Ryan Cotterell, Tanmoy Chakraborty, Yichao Zhou
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4515–4532
Language:
URL:
https://aclanthology.org/2021.naacl-main.357
DOI:
10.18653/v1/2021.naacl-main.357
Bibkey:
Cite (ACL):
Adithya V Ganesan, Matthew Matero, Aravind Reddy Ravula, Huy Vu, and H. Andrew Schwartz. 2021. Empirical Evaluation of Pre-trained Transformers for Human-Level NLP: The Role of Sample Size and Dimensionality. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4515–4532, Online. Association for Computational Linguistics.
Cite (Informal):
Empirical Evaluation of Pre-trained Transformers for Human-Level NLP: The Role of Sample Size and Dimensionality (V Ganesan et al., NAACL 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.naacl-main.357.pdf
Video:
 https://aclanthology.org/2021.naacl-main.357.mp4
Code
 adithya8/ContextualEmbeddingDR