Guided Dialogue Policy Learning without Adversarial Learning in the Loop

Ziming Li, Sungjin Lee, Baolin Peng, Jinchao Li, Julia Kiseleva, Maarten de Rijke, Shahin Shayandeh, Jianfeng Gao


Abstract
Reinforcement learning methods have emerged as a popular choice for training an efficient and effective dialogue policy. However, these methods suffer from sparse and unstable reward signals returned by a user simulator only when a dialogue finishes. Besides, the reward signal is manually designed by human experts, which requires domain knowledge. Recently, a number of adversarial learning methods have been proposed to learn the reward function together with the dialogue policy. However, to alternatively update the dialogue policy and the reward model on the fly, we are limited to policy-gradient-based algorithms, such as REINFORCE and PPO. Moreover, the alternating training of a dialogue agent and the reward model can easily get stuck in local optima or result in mode collapse. To overcome the listed issues, we propose to decompose the adversarial training into two steps. First, we train the discriminator with an auxiliary dialogue generator and then incorporate a derived reward model into a common reinforcement learning method to guide the dialogue policy learning. This approach is applicable to both on-policy and off-policy reinforcement learning methods. Based on our extensive experimentation, we can conclude the proposed method: (1) achieves a remarkable task success rate using both on-policy and off-policy reinforcement learning methods; and (2) has potential to transfer knowledge from existing domains to a new domain.
Anthology ID:
2020.findings-emnlp.209
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2020
Month:
November
Year:
2020
Address:
Online
Editors:
Trevor Cohn, Yulan He, Yang Liu
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
2308–2317
Language:
URL:
https://aclanthology.org/2020.findings-emnlp.209
DOI:
10.18653/v1/2020.findings-emnlp.209
Bibkey:
Cite (ACL):
Ziming Li, Sungjin Lee, Baolin Peng, Jinchao Li, Julia Kiseleva, Maarten de Rijke, Shahin Shayandeh, and Jianfeng Gao. 2020. Guided Dialogue Policy Learning without Adversarial Learning in the Loop. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 2308–2317, Online. Association for Computational Linguistics.
Cite (Informal):
Guided Dialogue Policy Learning without Adversarial Learning in the Loop (Li et al., Findings 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.findings-emnlp.209.pdf
Code
 cszmli/dp-without-adv