Reinforcement Learning for Turn-Taking Management in Incremental Spoken Dialogue Systems

Khouzaimi, Hatim; Laroche, Romain; Lefèvre, Fabrice

Reinforcement Learning for Turn-Taking Management in Incremental Spoken Dialogue Systems

Hatim Khouzaimi, Romain Laroche, Fabrice Lefèvre

Download BibTeX Cite

Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence. Pages 2831-2836.

Abstract

In this article, reinforcement learning is used to learn an optimal turn-taking strategy for vocal human-machine dialogue. The Orange Labs' Majordomo dialogue system, which allows the users to have conversations within a smart home, has been upgraded to an incremental version. First, a user simulator is built in order to generate a dialogue corpus which thereafter is used to optimise the turn-taking strategy from delayed rewards with the Fitted-Q reinforcement learning algorithm. Real users test and evaluate the new learnt strategy, versus a non-incremental and a handcrafted incremental strategies. The data-driven strategy is shown to significantly improve the task completion ratio and to be preferred by the users according to subjective metrics.

Hatim Khouzaimi, Romain Laroche, and Fabrice Lefèvre. 2016. Reinforcement Learning for Turn-Taking Management in Incremental Spoken Dialogue Systems. In Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-16. IJCAI, 2831–2836. https://ijcai.org/proceedings/2016/402

@inproceedings{ijcai2016p402,
  title     = {Reinforcement Learning for Turn-Taking Management in Incremental Spoken Dialogue Systems},
  author    = {Khouzaimi, Hatim and Laroche, Romain and Lefèvre, Fabrice},
  booktitle = {Proceedings of the Twenty-Fifth International Joint Conference on
               Artificial Intelligence, {IJCAI-16}},
  publisher = {AAAI Press / International Joint Conferences on Artificial Intelligence},
  editor    = {Subbarao Kambhampati},
  pages     = {2831--2836},
  year      = {2016},
  month     = {7},
  url       = {https://ijcai.org/proceedings/2016/402},
}

TY  - CPAPER
TI  - Reinforcement Learning for Turn-Taking Management in Incremental Spoken Dialogue Systems
AU  - Khouzaimi, Hatim
AU  - Laroche, Romain
AU  - Lefèvre, Fabrice
T2  - Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-16
PB  - AAAI Press / International Joint Conferences on Artificial Intelligence
PY  - 2016
SP  - 2831
EP  - 2836
UR  - https://ijcai.org/proceedings/2016/402
ER  -