168 search results - page 4 / 34 » Reinforcement Learning Algorithm for Partially Observable Ma... |
ECAI
2008
Springer
13 years 8 months ago
2008
Springer
Time is a crucial variable in planning and often requires special attention since it introduces a specific structure along with additional complexity, especially in the case of dec...
ECML
2003
Springer
13 years 11 months ago
2003
Springer
Due to the unavoidable fact that a robot’s sensors will be limited in some manner, it is entirely possible that it can find itself unable to distinguish between differing state...
ICRA
2007
IEEE
14 years 17 days ago
2007
IEEE
— We introduce the Oracular Partially Observable Markov Decision Process (OPOMDP), a type of POMDP in which the world produces no observations; instead there is an “oracle,” ...
CORR
2010
Springer
13 years 4 months ago
2010
Springer
We consider model-based reinforcement learning in finite Markov Decision Processes (MDPs), focussing on so-called optimistic strategies. Optimism is usually implemented by carryin...
AAAI
2011
12 years 6 months ago
2011
Recently, a number of researchers have proposed spectral algorithms for learning models of dynamical systems—for example, Hidden Markov Models (HMMs), Partially Observable Marko...