Using background knowledge to speed reinforcement learning in physical agents

Proceedings of the International Conference on Autonomous Agents

Volumn , Issue , 2001, Pages 254-261

  Shapiro, D ^a   Langley, P ^a   Shachter, R ^a  

^a Stanford University   (United States)

Author keywords

Action selection and planning; Adaptation and learning; Agent architectures; Hierarchical reinforcement learning

Indexed keywords

HIERARCHICAL SYSTEMS; KNOWLEDGE ACQUISITION; LEARNING ALGORITHMS; LEARNING SYSTEMS;

AGENT ARCHITECTURES; HIERARCHICAL REINFORCEMENT LEARNING;

MULTI AGENT SYSTEMS;

EID: 0034819986     PISSN: None     EISSN: None     Source Type: Conference Proceeding    
DOI: 10.1145/375735.376305     Document Type: Conference Paper

References (23)

1. A robust layered control system for a mobile robot
2. Feudal reinforcement learning
3. The MAXQ method for hierarchical reinforcement learning
4. State abstraction in MAXQ hierarchical reinforcement learning
5. A procedural logic
6. Reinforcement learning: A survey
7. Hierarchical learning in stochastic domains: Preliminary results
8. Chunking in Soar: The anatomy of a general learning mechanism
9. Teleoreactive programs for agent control
10. Reinforcement learning with hierarchies of machines
11. A preliminary analysis of the Soar architecture as a basis for general intelligence
12. Universal plans for reactive robots in unpredictable environments
13. Controlling physical agents through reactive logic programming
14. Convergence results for single-step on-policy reinforcement learning algorithms
15. Intra-option learning about temporally abstract actions
16. Integrating planning and learning: The PRODIGY architecture