Zijing Shi


2023

pdf bib
Self-imitation Learning for Action Generation in Text-based Games
Zijing Shi | Yunqiu Xu | Meng Fang | Ling Chen
Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics

In this work, we study reinforcement learning (RL) in solving text-based games. We address the challenge of combinatorial action space, by proposing a confidence-based self-imitation model to generate action candidates for the RL agent. Firstly, we leverage the self-imitation learning to rank and exploit past valuable trajectories to adapt a pre-trained language model (LM) towards a target game. Then, we devise a confidence-based strategy to measure the LM’s confidence with respect to a state, thus adaptively pruning the generated actions to yield a more compact set of action candidates. In multiple challenging games, our model demonstrates promising performance in comparison to the baselines.

pdf bib
CHBias: Bias Evaluation and Mitigation of Chinese Conversational Language Models
Jiaxu Zhao | Meng Fang | Zijing Shi | Yitong Li | Ling Chen | Mykola Pechenizkiy
Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

redWarning: This paper contains content that may be offensive or upsetting.Pretrained conversational agents have been exposed to safety issues, exhibiting a range of stereotypical human biases such as gender bias. However, there are still limited bias categories in current research, and most of them only focus on English. In this paper, we introduce a new Chinese dataset, CHBias, for bias evaluation and mitigation of Chinese conversational language models.Apart from those previous well-explored bias categories, CHBias includes under-explored bias categories, such as ageism and appearance biases, which received less attention. We evaluate two popular pretrained Chinese conversational models, CDial-GPT and EVA2.0, using CHBias. Furthermore, to mitigate different biases, we apply several debiasing methods to the Chinese pretrained models. Experimental results show that these Chinese pretrained models are potentially risky for generating texts that contain social biases, and debiasing methods using the proposed dataset can make response generation less biased while preserving the models’ conversational capabilities.