AGR: Reinforced Causal Agent-Guided Self-explaining Rationalization

Page view(s)
17
Checked on Oct 01, 2024
AGR: Reinforced Causal Agent-Guided Self-explaining Rationalization
Title:
AGR: Reinforced Causal Agent-Guided Self-explaining Rationalization
Journal Title:
Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
DOI:
Keywords:
Publication Date:
16 August 2024
Citation:
Yunxiao Zhao, Zhiqiang Wang, Xiaoli Li, Jiye Liang, and Ru Li. 2024. AGR: Reinforced Causal Agent-Guided Self-explaining Rationalization. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 510–518, Bangkok, Thailand. Association for Computational Linguistics.
Abstract:
Most existing rationalization approaches are susceptible to degeneration accumulation due to a lack of effective control over the learning direction of the model during training. To address this issue, we propose a novel approach AGR (Agent-Guided Rationalization), guiding the next action of the model based on its current training state. Specifically, we introduce causal intervention calculus to quantify the causal effects inherent during rationale training, and utilize reinforcement learning process to refine the learning bias of them. Furthermore, we pretrain an agent within this reinforced causal environment to guide the next step of the model. We theoretically demonstrate that a good model needs the desired guidance, and empirically show the effectiveness of our approach, outperforming existing state-of-the-art methods on BeerAdvocate and HotelReview datasets.
License type:
Attribution 4.0 International (CC BY 4.0)
Funding Info:
This work was supported by the National Natural Science Foundation of China (Nos.62376144, 62272285, 62076155) and the Science and Technology Co-operation and Exchange Special Project of Shanxi Province (No.202204041101016)
Description:
ISSN:
Nil
Files uploaded:

File Size Format Action
1611-agr-reinforced-causal-agent-gu.pdf 660.13 KB PDF Open