Addressing Sample Efficiency and Model-bias in Model-based Reinforcement Learning

Akhil S Anand*, Fares Abu-Dakka, Jan Tommy Gravdahl

*Tämän työn vastaava kirjoittaja

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussaConference article in proceedingsScientificvertaisarvioitu

2 Sitaatiot (Scopus)

Abstrakti

Model-based reinforcement learning promises to be an effective way to bring reinforcement learning to realworld robotic systems by offering a sample efficient learning approach compared to model-free reinforcement learning. However, model-based reinforcement learning approaches at present struggle to match the performance of model-free ones. This work attempts to fill this gap by improving the performance of model-based reinforcement learning while further improving its sample efficiency. To improve the sample efficiency, an exploration strategy is formulated which maximizes the information gain. The asymptotic performance is improved by compensating for the model-bias using a model-free critic. We have evaluated our proposed approach on four reinforcement learning benchmarking tasks in the openAI gym framework.
AlkuperäiskieliEnglanti
Otsikko21st IEEE International Conference on Machine Learning and Applications (IEEE ICMLA)
AlaotsikkoIEEE ICMLA 2023
ToimittajatM. Arif Wani, Mehmed Kantardzic, Vasile Palade, Daniel Neagu, Longzhi Yang, Kit-Yan Chan
KustantajaIEEE
Sivumäärä6
ISBN (elektroninen)978-1-6654-6283-9
DOI - pysyväislinkit
TilaJulkaistu - 23 maalisk. 2023
OKM-julkaisutyyppiA4 Artikkeli konferenssijulkaisussa
TapahtumaIEEE International Conference on Machine Learning and Applications - Nassau, Bahamasaaret
Kesto: 12 jouluk. 202214 jouluk. 2022
Konferenssinumero: 21

Conference

ConferenceIEEE International Conference on Machine Learning and Applications
LyhennettäICMLA
Maa/AlueBahamasaaret
KaupunkiNassau
Ajanjakso12/12/202214/12/2022

Sormenjälki

Sukella tutkimusaiheisiin 'Addressing Sample Efficiency and Model-bias in Model-based Reinforcement Learning'. Ne muodostavat yhdessä ainutlaatuisen sormenjäljen.

Siteeraa tätä