Safe Reinforcement Learning with Probabilistic Guarantees Satisfying Temporal Logic Specifications in Continuous Action Spaces.
Hanna KrasowskiPrithvi AkellaAaron D. AmesMatthias AlthoffPublished in: CDC (2023)
Keyphrases
- action space
- temporal logic
- reinforcement learning
- concurrent systems
- state space
- model checking
- transition systems
- markov decision processes
- bounded model checking
- reactive systems
- model checker
- continuous state
- real valued
- state and action spaces
- formal specification language
- continuous state spaces
- formal specification
- modal logic
- markov decision process
- finite state
- continuous action
- action selection
- state action
- bayesian networks
- stochastic processes
- dynamic programming
- function approximation
- function approximators
- reinforcement learning algorithms
- single agent
- model free
- verification method
- linear temporal logic
- control policies
- belief revision
- machine learning
- cooperative
- probabilistic model
- domain specific
- graphical models
- multiple agents
- formal verification
- policy iteration