Show simple item record

dc.rights.licenseCreative Commons "Este é um artigo publicado em acesso aberto sob uma licença Creative Commons (CC BY-NC-ND 4.0). Fonte: https://www.sciencedirect.com/science/article/pii/S095741742030244X?via%3Dihub. Acesso em: 23 nov. 2021.
dc.contributor.authorGLATT, RUBEN
dc.contributor.authorSILVA, FELIPE LENO DA
dc.contributor.authorReinaldo Bianchi
dc.contributor.authorCOSTA, ANNA HELENA REALI
dc.date.accessioned2021-11-24T00:35:41Z
dc.date.accessioned2023-05-03T20:33:19Z
dc.date.available2021-11-24T00:35:41Z
dc.date.available2023-05-03T20:33:19Z
dc.date.issued2020-10-15
dc.identifier.citationGLATT, R.; S., F. L. DA; BIANCHI, R. A. DA C.; COSTA, A. H R. DECAF: Deep Case-based Policy Inference for Knowledge Transfer in Reinforcement Learning. EXPERT SYSTEMS WITH APPLICATIONS, v. 1, p. 113420, 2020.
dc.identifier.issn0957-4174
dc.identifier.urihttps://hdl.handle.net/20.500.12032/88596
dc.description.abstractHaving the ability to solve increasingly complex problems using Reinforcement Learning (RL) has prompted researchers to start developing a greater interest in systematic approaches to retain and reuse knowledge over a variety of tasks. With Case-based Reasoning (CBR) there exists a general methodology that provides a framework for knowledge transfer which has been underrepresented in the RL literature so far. We for- mulate a terminology for the CBR framework targeted towards RL researchers with the goal of facilitating communication between the respective research communities. Based on this framework, we propose the Deep Case-based Policy Inference (DECAF) algorithm to accelerate learning by building a library of cases and reusing them if they are similar to a new task when training a new policy. DECAF guides the train- ing by dynamically selecting and blending policies according to their usefulness for the current target task, reusing previously learned policies for a more effective exploration but still enabling the adaptation to particularities of the new task. We show an empirical evaluation in the Atari game playing domain depicting the benefits of our algorithm with regards to sample efficiency, robustness against negative transfer, and performance increase when compared to state-of-the-art methods.
dc.relation.ispartofEXPERT SYSTEMS WITH APPLICATIONS
dc.rightsAcesso Restrito
dc.subjectDeep Reinforcement Learning
dc.subjectCase-based Reasoning
dc.subjectTransfer Learning
dc.subjectKnowledge discovery
dc.subjectKnowledge management
dc.subjectNeural networks
dc.titleDECAF: Deep Case-based Policy Inference for Knowledge Transfer in Reinforcement Learningpt_BR
dc.typeArtigopt_BR


Files in this item

FilesSizeFormatView

This item appears in the following Collection(s)

Show simple item record


© AUSJAL 2022

Asociación de Universidades Confiadas a la Compañía de Jesús en América Latina, AUSJAL
Av. Santa Teresa de Jesús Edif. Cerpe, Piso 2, Oficina AUSJAL Urb.
La Castellana, Chacao (1060) Caracas - Venezuela
Tel/Fax (+58-212)-266-13-41 /(+58-212)-266-85-62

Nuestras redes sociales

facebook Facebook

twitter Twitter

youtube Youtube

Asociaciones Jesuitas en el mundo
Ausjal en el mundo AJCU AUSJAL JESAM JCEP JCS JCAP