Understanding Finite-State Representations of Recurrent Policy Networks.
Mohamad H. DaneshAnurag KoulAlan FernSaeed KhorramPublished in: CoRR (2020)
Keyphrases
- finite state
- optimal policy
- markov chain
- partially observable markov decision processes
- policy iteration algorithm
- markov decision processes
- average cost
- model checking
- policy iteration
- state space
- reinforcement learning
- state dependent
- recurrent neural networks
- decision problems
- markov decision process
- transition systems
- dynamic programming
- tree automata
- continuous state
- multistage
- probabilistic model
- stationary policies
- function approximation
- infinite horizon
- search algorithm
- control policies
- information retrieval