Tight Performance Guarantees of Imitator Policies with Continuous Actions.
Davide MaranAlberto Maria MetelliMarcello RestelliPublished in: AAAI (2023)
Keyphrases
- continuous action
- initial state
- action space
- optimal policy
- reward function
- lower bound
- decision processes
- selective perception
- upper bound
- decision theoretic
- macro actions
- policy search
- plan recognition
- action selection
- temporally extended
- markov decision process
- neural network
- goal directed
- piecewise linear
- human activities
- markov decision processes
- situation calculus
- state transitions
- reasoning about actions
- decision theoretic planning
- multiagent reinforcement learning
- worst case
- reinforcement learning
- decision making
- computer vision