Content Masked Loss: Human-Like Brush Stroke Planning in a Reinforcement Learning Painting Agent.
Peter SchaldenbrandJean OhPublished in: CoRR (2020)
Keyphrases
- reinforcement learning
- action selection
- multi agent
- partially observable
- agent architecture
- reactive agents
- single agent
- multi agent systems
- stochastic domains
- learning agent
- bdi agents
- state abstraction
- decision theoretic
- external events
- multimedia
- learning capabilities
- virtual agents
- learning process
- intelligent agents
- semantic web service composition
- metadata
- planning process
- agent receives
- state space
- agent model
- autonomous learning
- function approximation
- learning agents
- reward shaping
- reward function
- plan execution
- information gathering
- action space
- multiple agents
- world model
- chinese characters
- software agents
- multi agent environments
- autonomous agents
- decision making
- dynamic programming
- optimal policy
- heuristic search
- agent learns
- markov decision processes
- blocks world
- goal directed
- machine learning
- markov decision process