Humans are not Boltzmann Distributions: Challenges and Opportunities for Modelling Human Feedback and Interaction in Reinforcement Learning.
David LindnerMennatallah El-AssadyPublished in: CoRR (2022)
Keyphrases
- human operators
- reinforcement learning
- human communication
- human human interaction
- human behavior
- human interaction
- human subjects
- user engagement
- human users
- human robot
- human interactions
- lessons learned
- human agent
- motor skills
- function approximation
- sensory inputs
- artificial agents
- body movements
- artificial systems
- human computer interaction
- real world
- learning algorithm
- human teacher
- human experts
- probability distribution
- human centric
- markov decision processes
- human robot interaction
- turing test
- user interaction
- conversational agent
- random variables
- state space
- model free
- virtual characters
- virtual agents
- human intelligence
- temporal difference
- relevance feedback
- nonverbal communication
- visual stimuli
- learning process
- optimal policy
- cognitive process
- human activities
- human cognition
- action selection
- physical world