Skip to Main content Skip to Navigation
Conference papers

Améliorer un agent conversationnel : prendre en compte à la volée des retours utilisateurs

Maxime Arens 1, 2
Abstract : We present an approach to improve the relevance of a conversational question answering systemby leveraging previous user feedback. A dialog system deployed in contact of users can take intoaccounts feedbacks to improve the relevance of its answers. Question answering systems usuallywork through models matching a question with one or multiple answers. Here we consider the casewhere the model matches a question to a list of answers scored by relevance. A classical approach ofconsidering user feedback is to augment the training data used to retrain the matching model. Herewe suggest a different approach, impacting answers scores, by considering “on the fly” the feedbacks :between when the user asks a new question and when the system responds.
Document type :
Conference papers
Complete list of metadata

https://hal.archives-ouvertes.fr/hal-03265903
Contributor : Yannick Parmentier Connect in order to contact the contributor
Submitted on : Wednesday, June 23, 2021 - 11:50:26 PM
Last modification on : Tuesday, October 19, 2021 - 2:24:25 PM
Long-term archiving on: : Friday, September 24, 2021 - 7:19:03 PM

File

54.pdf
Publisher files allowed on an open archive

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

  • HAL Id : hal-03265903, version 1

Citation

Maxime Arens. Améliorer un agent conversationnel : prendre en compte à la volée des retours utilisateurs. 23e Rencontre des Étudiants Chercheurs en Informatique pour le Traitement Automatique des Langues @ 28e Conférence sur le Traitement Automatique des Langues Naturelles (RECITAL @ TALN 2021), ATALA, Jun 2021, mode virtuel, France. pp.9-21. ⟨hal-03265903⟩

Share

Metrics

Record views

85

Files downloads

66