TopicBERT for Energy Efficient Document Classification

Yatin Chaudhary, Pankaj Gupta, Khushbu Saxena, Vivek Kulkarni, Thomas Runkler, Hinrich Schütze


Abstract
Prior research notes that BERT’s computational cost grows quadratically with sequence length thus leading to longer training times, higher GPU memory constraints and carbon emissions. While recent work seeks to address these scalability issues at pre-training, these issues are also prominent in fine-tuning especially for long sequence tasks like document classification. Our work thus focuses on optimizing the computational cost of fine-tuning for document classification. We achieve this by complementary learning of both topic and language models in a unified framework, named TopicBERT. This significantly reduces the number of self-attention operations – a main performance bottleneck. Consequently, our model achieves a 1.4x ( 40%) speedup with 40% reduction in CO2 emission while retaining 99.9% performance over 5 datasets.
Anthology ID:
2020.findings-emnlp.152
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2020
Month:
November
Year:
2020
Address:
Online
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1682–1690
Language:
URL:
https://aclanthology.org/2020.findings-emnlp.152
DOI:
10.18653/v1/2020.findings-emnlp.152
Bibkey:
Cite (ACL):
Yatin Chaudhary, Pankaj Gupta, Khushbu Saxena, Vivek Kulkarni, Thomas Runkler, and Hinrich Schütze. 2020. TopicBERT for Energy Efficient Document Classification. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 1682–1690, Online. Association for Computational Linguistics.
Cite (Informal):
TopicBERT for Energy Efficient Document Classification (Chaudhary et al., Findings 2020)
Copy Citation:
PDF:
https://preview.aclanthology.org/auto-file-uploads/2020.findings-emnlp.152.pdf
Video:
 https://slideslive.com/38940122
Code
 YatinChaudhary/TopicBERT