Where to Add Actions in Human-in-the-Loop Reinforcement Learning.
Travis MandelYun-En LiuEmma BrunskillZoran PopovicPublished in: AAAI (2017)
Keyphrases
- reinforcement learning
- behavioural cloning
- action selection
- perceptual aliasing
- human activities
- function approximation
- multiagent reinforcement learning
- partial observability
- learning algorithm
- action space
- human interaction
- human beings
- temporal difference
- video sequences
- hand movements
- sensory inputs
- partially observable
- human operators
- reward function
- human experts
- markov decision processes
- neural network
- situation calculus
- reinforcement learning methods
- activity recognition
- optimal policy
- state and action spaces
- state space
- partially observable domains
- machine learning