Mission-driven Exploration for Accelerated Deep Reinforcement Learning with Temporal Logic Task Specifications.
Jun WangHosein HasanbeigKaiyuan TanZihe SunYiannis KantarosPublished in: CoRR (2023)
Keyphrases
- temporal logic
- reinforcement learning
- concurrent systems
- transition systems
- model checking
- reactive systems
- bounded model checking
- model checker
- active exploration
- formal specification language
- action selection
- formal specification
- modal logic
- verification method
- satisfiability problem
- function approximation
- state space
- multi agent
- computation tree logic
- linear temporal logic
- belief revision
- temporal knowledge
- markov decision processes
- mazurkiewicz traces
- dynamic constraints
- predicate logic
- formal verification
- optimal policy
- dynamic programming
- search strategies