Carregant...
Carregant...

Vés al contingut (premeu Retorn)

Competitive function approximation for reinforcement learning

Autor
Agostini, A.; Celaya, E.
Tipus d'activitat
Document cientificotècnic
Data
2014
Codi
IRI-TR-14-05
Repositori
http://hdl.handle.net/2117/28454 Obrir en finestra nova
Resum
The application of reinforcement learning to problems with continuous domains requires representing the value function by means of function approximation. We identify two aspects of reinforcement learning that make the function approximation process hard: non-stationarity of the target function and biased sampling. Non-stationarity is the result of the bootstrapping nature of dynamic programming where the value function is estimated using its current approximation. Biased sampling occurs when so...
Citació
Agostini, A.; Celaya, E. "Competitive function approximation for reinforcement learning". 2014.
Paraules clau
Gaussian mixture model, competitive strategy, learning (artificial intelligence), reinforcement learning
Grup de recerca
KRD - Cinemàtica i Disseny de Robots

Participants