Explaining a Deep Reinforcement Learning Docking Agent Using Linear Model Trees with User Adapted Visualization.
Vilde B. GjærumInga StrümkeOle Andreas AlsosAnastasios M. LekkasPublished in: CoRR (2022)
Keyphrases
- linear model
- reinforcement learning
- regression trees
- multi agent
- least squares
- linear models
- regression model
- agent receives
- action selection
- autonomous agents
- multi agent systems
- interface agent
- multi agent environments
- learning capabilities
- optimal policy
- learning agents
- reward function
- multiagent systems
- human users
- intelligent agents
- user interface
- linear transformation
- nonlinear models
- partially observable
- closed loop
- learning algorithm
- markov decision processes
- markov decision process
- feature selection
- learning agent
- model free
- exploration strategy
- decision trees
- function approximation
- machine learning