Reinforcement Learning with Limited Reinforcement: Using Bayes Risk for Active Learning in POMDPs.
Finale DoshiJoelle PineauNicholas RoyPublished in: ISAIM (2008)
Keyphrases
- reinforcement learning
- bayes risk
- active learning
- upper and lower bounds
- transfer learning
- learning algorithm
- machine learning
- partially observable markov decision processes
- loss function
- state space
- supervised learning
- learning process
- continuous state
- reproducing kernel hilbert space
- posterior probability
- random sampling
- distortion measure
- kernel methods
- sample complexity
- support vector
- optimal policy
- semi supervised
- learning problems
- conditional probabilities
- real valued
- semi supervised learning
- squared error
- labeled data