Robot Representation and Reasoning with Knowledge from Reinforcement Learning
This work addresses the challenge of combining logical reasoning with experiential learning for robots, offering a hybrid approach that is incremental in bridging two AI paradigms.
The paper tackles the problem of integrating declarative knowledge representation and reasoning with reinforcement learning to enable agents to reason with knowledge and learn from interactions, resulting in significant improvements over existing model-based RL methods in robot tasks like dialog, navigation, and delivery.
Reinforcement learning (RL) agents aim at learning by interacting with an environment, and are not designed for representing or reasoning with declarative knowledge. Knowledge representation and reasoning (KRR) paradigms are strong in declarative KRR tasks, but are ill-equipped to learn from such experiences. In this work, we integrate logical-probabilistic KRR with model-based RL, enabling agents to simultaneously reason with declarative knowledge and learn from interaction experiences. The knowledge from humans and RL is unified and used for dynamically computing task-specific planning models under potentially new environments. Experiments were conducted using a mobile robot working on dialog, navigation, and delivery tasks. Results show significant improvements, in comparison to existing model-based RL methods.