Optimiser l'adaptation en ligne d'un module de compréhension de la parole avec un algorithme de bandit contre un adversaire

Abstract : Adversarial bandit for optimising online active learning of spoken language understanding Many speech understanding modules have in common to be probabilistic and to rely on machine learning algorithms to train their models from large amount of data. The difficulty remains in the cost of collecting such data and the time for updating an existing model to a new domain. In this paper, we propose to drive an online adaptive process with a policy learnt using the Adversarial Bandit algorithm. We showthat this proposition can optimally balance the cost of gathering valuable user feedbacks and the overall performance of the spoken language understanding module after its update.
Document type :
Conference papers
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-02063583
Contributor : Bassam Jabaian <>
Submitted on : Monday, March 11, 2019 - 12:20:50 PM
Last modification on : Wednesday, May 15, 2019 - 10:12:14 AM

Identifiers

  • HAL Id : hal-02063583, version 1

Collections

Citation

Emmanuel Ferreira, Alexandre Reiffers-Masson, Bassam Jabaian, Fabrice Lefevre. Optimiser l'adaptation en ligne d'un module de compréhension de la parole avec un algorithme de bandit contre un adversaire. JEP, 2016, Paris, France. ⟨hal-02063583⟩

Share

Metrics

Record views

17