Skip to Main content Skip to Navigation
Conference papers

Décodage guidé par un discriminateur avec le Monte Carlo Tree Search pour la génération de texte contrainte

Antoine Chaffin 1, 2 Vincent Claveau 2 Ewa Kijak 2 
2 LinkMedia - Creating and exploiting explicit links between multimedia fragments
Inria Rennes – Bretagne Atlantique , IRISA-D6 - SIGNAL, IMAGE ET LANGAGE
Abstract : In this paper, we explore how to control text generation at decoding time to satisfy certain constraints (eg. being non-toxic, conveying certain emotions...) without fine-tuning the language model. Precisely, we formalize constrained generation as a tree exploration process guided by a discriminator that indicates how well the associated sequence respects the constraint. We propose several original methods to search this generation tree, notably the Monte Carlo Tree Search (MCTS) which provides theoretical guarantees on the search efficiency.Through 3 tasks and 2 languages, we show that discriminator-guided MCTS decoding achieves state-of-the-art results without having to tune the language model. We also demonstrate that other proposed decoding methods based on re-ranking can be really effective when diversity among the generated propositions is encouraged.
Document type :
Conference papers
Complete list of metadata
Contributor : Yannick Parmentier Connect in order to contact the contributor
Submitted on : Friday, June 24, 2022 - 4:41:34 PM
Last modification on : Friday, August 5, 2022 - 2:54:52 PM
Long-term archiving on: : Sunday, September 25, 2022 - 9:34:04 PM


Publisher files allowed on an open archive


  • HAL Id : hal-03701490, version 1


Antoine Chaffin, Vincent Claveau, Ewa Kijak. Décodage guidé par un discriminateur avec le Monte Carlo Tree Search pour la génération de texte contrainte. TALN 2022 - 29e conférence sur le Traitement Automatique des Langues Naturelles, Jun 2022, Avignon, France. pp.27-41. ⟨hal-03701490⟩



Record views


Files downloads