Boosting Active Learning to Optimality: A Tractable Monte-Carlo, Billiard-Based Algorithm.
Philippe RoletMichèle SebagOlivier TeytaudPublished in: ECML/PKDD (2) (2009)
Keyphrases
- monte carlo
- learning algorithm
- active learning
- optimal solution
- np hard
- monte carlo simulation
- computational complexity
- importance sampling
- markov chain
- matrix inversion
- simulation study
- simulated annealing
- dynamic programming
- search space
- probabilistic model
- game tree search
- stochastic approximation
- worst case
- markov chain monte carlo
- kalman filter
- upper bound
- objective function
- decision trees