Transfer of Deep Reactive Policies for MDP Planning.
Aniket (Nick) BajpaiSankalp Garg MausamPublished in: NeurIPS (2018)
Keyphrases
- optimal policy
- markov decision problems
- markov decision process
- partially observable markov decision processes
- markov decision processes
- reactive planning
- macro actions
- reinforcement learning
- state space
- partially observable
- stochastic domains
- planning under uncertainty
- planning problems
- initial state
- reactive agents
- probabilistic planning
- linear programming
- decision problems
- utility function
- decision theoretic
- finite state
- decision theoretic planning
- factored markov decision processes
- bayesian reinforcement learning
- admissible heuristics
- policy iteration
- reward function
- decision processes
- average cost
- ai planning
- linear program
- domain independent
- blocks world
- planning systems
- policy search
- reinforcement learning problems
- infinite horizon
- dynamical systems
- predictive state representations
- transfer learning
- dynamic programming