Eliminating Sentiment Bias for Aspect-Level Sentiment Classification with Unsupervised Opinion Extraction

Bo Wang, Tao Shen, Guodong Long, Tianyi Zhou, Yi Chang


Abstract
Aspect-level sentiment classification (ALSC) aims at identifying the sentiment polarity of a specified aspect in a sentence. ALSC is a practical setting in aspect-based sentiment analysis due to no opinion term labeling needed, but it fails to interpret why a sentiment polarity is derived for the aspect. To address this problem, recent works fine-tune pre-trained Transformer encoders for ALSC to extract an aspect-centric dependency tree that can locate the opinion words. However, the induced opinion words only provide an intuitive cue far below human-level interpretability. Besides, the pre-trained encoder tends to internalize an aspect’s intrinsic sentiment, causing sentiment bias and thus affecting model performance. In this paper, we propose a span-based anti-bias aspect representation learning framework. It first eliminates the sentiment bias in the aspect embedding by adversarial learning against aspects’ prior sentiment. Then, it aligns the distilled opinion candidates with the aspect by span-based dependency modeling to highlight the interpretable opinion terms. Our method achieves new state-of-the-art performance on five benchmarks, with the capability of unsupervised opinion extraction.
Anthology ID:
2021.findings-emnlp.258
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2021
Month:
November
Year:
2021
Address:
Punta Cana, Dominican Republic
Editors:
Marie-Francine Moens, Xuanjing Huang, Lucia Specia, Scott Wen-tau Yih
Venue:
Findings
SIG:
SIGDAT
Publisher:
Association for Computational Linguistics
Note:
Pages:
3002–3012
Language:
URL:
https://aclanthology.org/2021.findings-emnlp.258
DOI:
10.18653/v1/2021.findings-emnlp.258
Bibkey:
Cite (ACL):
Bo Wang, Tao Shen, Guodong Long, Tianyi Zhou, and Yi Chang. 2021. Eliminating Sentiment Bias for Aspect-Level Sentiment Classification with Unsupervised Opinion Extraction. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 3002–3012, Punta Cana, Dominican Republic. Association for Computational Linguistics.
Cite (Informal):
Eliminating Sentiment Bias for Aspect-Level Sentiment Classification with Unsupervised Opinion Extraction (Wang et al., Findings 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.findings-emnlp.258.pdf
Video:
 https://aclanthology.org/2021.findings-emnlp.258.mp4
Code
 wangbo9719/sarl_absa
Data
ASTE