A novel orthogonal NMF-based belief compression for POMDPs.
Xin LiWilliam Kwok-Wai CheungJiming LiuZhili WuPublished in: ICML (2007)
Keyphrases
- belief state
- point based value iteration
- negative matrix factorization
- belief space
- nonnegative matrix factorization
- state space
- image compression
- compression scheme
- data compression
- reinforcement learning
- belief revision
- compression ratio
- partially observable markov decision processes
- compression algorithm
- partially observable
- matrix factorization
- belief functions
- frequent sets
- dynamic programming
- feature extraction
- continuous state
- pairwise
- planning under uncertainty
- data matrix
- probability theory
- markov decision processes
- finite state
- document clustering