Content Masked Loss: Human-Like Brush Stroke Planning in a Reinforcement Learning Painting Agent.
Peter SchaldenbrandJean OhPublished in: AAAI (2021)
Keyphrases
- reinforcement learning
- action selection
- multi agent
- partially observable
- agent architecture
- single agent
- reactive agents
- stochastic domains
- reward shaping
- decision theoretic
- multi agent systems
- external events
- intelligent agents
- autonomous agents
- learning agent
- state space
- semantic web service composition
- autonomous learning
- bdi agents
- decision making
- multimedia
- metadata
- multiagent systems
- planning process
- planning problems
- agent behavior
- world model
- function approximation
- software agents
- mobile agents
- virtual agents
- plan execution
- reward function
- goal directed
- information gathering
- learning capabilities
- temporal difference
- agent technology
- learning agents
- model free
- agent receives
- state abstraction
- multi agent environments
- domain independent
- optimal policy
- machine learning
- agent learns
- partially observable markov decision process
- markov decision problems
- chinese characters
- multiple agents
- human users