Asking for Knowledge : Training RL Agents to Query External Knowledge Using Language - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

Asking for Knowledge : Training RL Agents to Query External Knowledge Using Language

Résumé

To solve difficult tasks, humans ask questions to acquire knowledge from external sources. In contrast, classical reinforcement learning agents lack such an ability and often resort to exploratory behavior. This is exacerbated as few present-day environments support querying for knowledge. In order to study how agents can be taught to query external knowledge via language, we first introduce two new environments: the grid-world-based Q-BabyAI and the text-based Q-TextWorld. In addition to physical interactions, an agent can query an external knowledge source specialized for these environments to gather information. Second, we propose the 'Asking for Knowledge' (AFK) agent, which learns to generate language commands to query for meaningful knowledge that helps solve the tasks. AFK leverages a non-parametric memory, a pointer mechanism and an episodic exploration bonus to tackle (1) irrelevant information, (2) a large query language space, (3) delayed reward for making meaningful queries. Extensive experiments demonstrate that the AFK agent outperforms recent baselines on the challenging Q-BabyAI and Q-TextWorld environments. The code of the environments and agents are available at https://ioujenliu.github.io/AFK.
Fichier principal
Vignette du fichier
liu22t.pdf (4.07 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03897379 , version 1 (13-12-2022)

Identifiants

  • HAL Id : hal-03897379 , version 1

Citer

Iou-Jen Liu, Xingdi Yuan, Marc-Alexandre Côté, Pierre-Yves Oudeyer, Alexander G Schwing. Asking for Knowledge : Training RL Agents to Query External Knowledge Using Language. ICML 2022 - 39th International Conference on Machine Learning, Jul 2022, Baltimore, United States. ⟨hal-03897379⟩

Collections

INRIA INRIA2
18 Consultations
17 Téléchargements

Partager

Gmail Facebook X LinkedIn More