Weakly Supervised Text Classification using Supervision Signals from a Language Model

Ziqian Zeng, Weimin Ni, Tianqing Fang, Xiang Li, Xinran Zhao, Yangqiu Song


Abstract
Solving text classification in a weakly supervised manner is important for real-world applications where human annotations are scarce. In this paper, we propose to query a masked language model with cloze style prompts to obtain supervision signals. We design a prompt which combines the document itself and “this article is talking about [MASK].” A masked language model can generate words for the [MASK] token. The generated words which summarize the content of a document can be utilized as supervision signals. We propose a latent variable model to learn a word distribution learner which associates generated words to pre-defined categories and a document classifier simultaneously without using any annotated data. Evaluation on three datasets, AGNews, 20Newsgroups, and UCINews, shows that our method can outperform baselines by 2%, 4%, and 3%.
Anthology ID:
2022.findings-naacl.176
Volume:
Findings of the Association for Computational Linguistics: NAACL 2022
Month:
July
Year:
2022
Address:
Seattle, United States
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2295–2305
Language:
URL:
https://aclanthology.org/2022.findings-naacl.176
DOI:
10.18653/v1/2022.findings-naacl.176
Bibkey:
Cite (ACL):
Ziqian Zeng, Weimin Ni, Tianqing Fang, Xiang Li, Xinran Zhao, and Yangqiu Song. 2022. Weakly Supervised Text Classification using Supervision Signals from a Language Model. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 2295–2305, Seattle, United States. Association for Computational Linguistics.
Cite (Informal):
Weakly Supervised Text Classification using Supervision Signals from a Language Model (Zeng et al., Findings 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingestion-script-update/2022.findings-naacl.176.pdf
Video:
 https://preview.aclanthology.org/ingestion-script-update/2022.findings-naacl.176.mp4
Code
 hkust-knowcomp/wddc
Data
AG News