116
Voted
NIPS
2000
15 years 4 months ago
2000
The problem of reinforcement learning in large factored Markov decision processes is explored. The Q-value of a state-action pair is approximated by the free energy of a product o...