SlateQ: A Tractable Decomposition for Reinforcement Learning with Recommendation Sets.
Eugene IeVihan JainJing WangSanmit NarvekarRitesh AgarwalRui WuHeng-Tze ChengTushar ChandraCraig BoutilierPublished in: IJCAI (2019)
Keyphrases
- reinforcement learning
- recommender systems
- version spaces
- function approximation
- personalized recommendation
- optimal policy
- learning process
- collaborative filtering
- markov decision processes
- recommendation systems
- model free
- robotic control
- hypertree decomposition
- definite clause
- decomposition methods
- np complete
- computational complexity
- learning algorithm
- reinforcement learning algorithms
- decomposition method
- supervised learning
- dynamic programming
- machine learning