English

Meta Dialogue Policy Learning

Computation and Language 2020-06-05 v1 Machine Learning

Abstract

Dialog policy determines the next-step actions for agents and hence is central to a dialogue system. However, when migrated to novel domains with little data, a policy model can fail to adapt due to insufficient interactions with the new environment. We propose Deep Transferable Q-Network (DTQN) to utilize shareable low-level signals between domains, such as dialogue acts and slots. We decompose the state and action representation space into feature subspaces corresponding to these low-level components to facilitate cross-domain knowledge transfer. Furthermore, we embed DTQN in a meta-learning framework and introduce Meta-DTQN with a dual-replay mechanism to enable effective off-policy training and adaptation. In experiments, our model outperforms baseline models in terms of both success rate and dialogue efficiency on the multi-domain dialogue dataset MultiWOZ 2.0.

Keywords

Cite

@article{arxiv.2006.02588,
  title  = {Meta Dialogue Policy Learning},
  author = {Yumo Xu and Chenguang Zhu and Baolin Peng and Michael Zeng},
  journal= {arXiv preprint arXiv:2006.02588},
  year   = {2020}
}

Comments

10 pages, 3 figures

R2 v1 2026-06-23T16:02:36.114Z