arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2004.14507cs.LGcs.AIcs.CL
收藏

用于神经响应生成的反事实非策略训练

Counterfactual Off-Policy Training for Neural Response Generation

Harbin Institute of Technology(哈尔滨工业大学);University of California, Santa Barbara(加利福尼亚大学圣巴巴拉分校)

Qingfu Zhu, Weinan Zhang, Ting Liu, William Yang Wang

摘要:

开放域对话生成由于潜在响应的巨大规模而遭受数据不足的问题。在本文中,我们建议通过反事实推理来探索潜在的反应。给定观察到的响应,反事实推理模型自动推断可能已经采取的替代策略的结果。事后合成的反事实反应比从头开始合成的反应质量更高。对抗性学习框架下的反事实反应训练有助于探索潜在反应空间的高回报区域。对DailyDialog数据集的实证研究表明,我们的方法显著优于HRED模型和传统的对抗性学习方法。

英文摘要:

Open-domain dialogue generation suffers from the data insufficiency problem due to the vast size of potential responses. In this paper, we propose to explore potential responses by counterfactual reasoning. Given an observed response, the counterfactual reasoning model automatically infers the outcome of an alternative policy that could have been taken. The resulting counterfactual response synthesized in hindsight is of higher quality than the response synthesized from scratch. Training on the counterfactual responses under the adversarial learning framework helps to explore the high-reward area of the potential response space. An empirical study on the DailyDialog dataset shows that our approach significantly outperforms the HRED model as well as the conventional adversarial learning approaches.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑