246 search results - page 1 / 50 » Approximate Predictive Representations of Partially Observab... |
ICML
2010
IEEE
13 years 2 months ago
2010
IEEE
We provide a novel view of learning an approximate model of a partially observable environment from data and present a simple implemenf the idea. The learned model abstracts away ...
ICMLA
2008
13 years 6 months ago
2008
Learning by imitation has shown to be a powerful paradigm for automated learning in autonomous robots. This paper presents a general framework of learning by imitation for stochas...
AAAI
1996
1996
Computing Optimal Policies for Partially Observable Decision Processes Using Compact Representations
13 years 6 months ago
: Partially-observable Markov decision processes provide a very general model for decision-theoretic planning problems, allowing the trade-offs between various courses of actions t...
AAAI
2011
12 years 4 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...
ATAL
2007
Springer
13 years 8 months ago
2007
Springer
Models of agent-environment interaction that use predictive state representations (PSRs) have mainly focused on the case of discrete observations and actions. The theory of discre...