Relational Reinforcement Learning in Infinite Mario

2012/02/28 by Shiwali Mohan, John E. Laird, Mohan, Shiwali +1
Computer Science · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #cs.AI

paper · pdf · doi:10.48550/arxiv.1202.6386

arxiv created 2012/02/28 · arxiv updated 2012/03/01

Abstract

Relational representations in reinforcement learning allow for the use of structural information like the presence of objects and relationships between them in the description of value functions. Through this paper, we show that such representations allow for the inclusion of background knowledge that qualitatively describes a state and can be used to design agents that demonstrate learning behavior in domains with large state and actions spaces such as computer games.

Related

• Learning Rules from Rewards
• Relational Deep Reinforcement Learning
• Deep Explainable Relational Reinforcement Learning: A Neuro-Symbolic Approach
• Symbolic Relational Deep Reinforcement Learning based on Graph Neural Networks and Autoregressive Policy Decomposition
• Human-Allied Relational Reinforcement Learning