Text Editing as Imitation Game

Ning Shi, Bin Tang, Bo Yuan, Longtao Huang, Yewen Pu, Jie Fu, Zhouhan Lin


Abstract
Text editing, such as grammatical error correction, arises naturally from imperfect textual data. Recent works frame text editing as a multi-round sequence tagging task, where operations – such as insertion and substitution – are represented as a sequence of tags. While achieving good results, this encoding is limited in flexibility as all actions are bound to token-level tags. In this work, we reformulate text editing as an imitation game using behavioral cloning. Specifically, we convert conventional sequence-to-sequence data into state-to-action demonstrations, where the action space can be as flexible as needed. Instead of generating the actions one at a time, we introduce a dual decoders structure to parallel the decoding while retaining the dependencies between action tokens, coupled with trajectory augmentation to alleviate the distribution shift that imitation learning often suffers. In experiments on a suite of Arithmetic Equation benchmarks, our model consistently outperforms the autoregressive baselines in terms of performance, efficiency, and robustness. We hope our findings will shed light on future studies in reinforcement learning applying sequence-level action generation to natural language processing.
Anthology ID:
2022.findings-emnlp.114
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2022
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Editors:
Yoav Goldberg, Zornitsa Kozareva, Yue Zhang
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1583–1594
Language:
URL:
https://aclanthology.org/2022.findings-emnlp.114
DOI:
10.18653/v1/2022.findings-emnlp.114
Bibkey:
Cite (ACL):
Ning Shi, Bin Tang, Bo Yuan, Longtao Huang, Yewen Pu, Jie Fu, and Zhouhan Lin. 2022. Text Editing as Imitation Game. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 1583–1594, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Text Editing as Imitation Game (Shi et al., Findings 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.findings-emnlp.114.pdf
Video:
 https://aclanthology.org/2022.findings-emnlp.114.mp4