Online learning and transfer for user adaptation in dialogue systems - Archive ouverte HAL Access content directly
Conference Papers Year : 2017

Online learning and transfer for user adaptation in dialogue systems

Nicolas Carrara
  • Function : Author
  • PersonId : 1012066
Romain Laroche
  • Function : Author
  • PersonId : 1012067
Olivier Pietquin

Abstract

We address the problem of user adaptation in Spoken Dialogue Systems. The goal is to quickly adapt online to a new user given a large amount of dialogues collected with other users. Previous works using Transfer for Reinforcement Learning tackled this problem when the number of source users remains limited. In this paper, we overcome this constraint by clustering the source users: each user cluster, represented by its centroid, is used as a potential source in the state-of-the-art Transfer Reinforcement Learning algorithm. Our benchmark compares several clustering approaches , including one based on a novel metric. All experiments are led on a negotiation dialogue task, and their results show significant improvements over baselines.
Fichier principal
Vignette du fichier
camera-ready.pdf (588.8 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01557775 , version 1 (06-07-2017)

Identifiers

  • HAL Id : hal-01557775 , version 1

Cite

Nicolas Carrara, Romain Laroche, Olivier Pietquin. Online learning and transfer for user adaptation in dialogue systems. SIGDIAL/SEMDIAL joint special session on negotiation dialog 2017, Aug 2017, Saarbrücken, Germany. ⟨hal-01557775⟩
498 View
316 Download

Share

Gmail Facebook X LinkedIn More