Using the event calculus to integrate planning and learning in an intelligent autonomous agent
Gunther Sablon, Maurice Bruynooghe · Lirias · 1994
An autonomous agent architecture, in which Machine Learning and Planning are integrated, is presented. The agent's chosen knowledge representation is the Event Calculus, which can be expressed in Horn clause logic. The paper shows that the logical approach of the Event Calculus is suitable for integrating planning and learning in the autonomous agent: an existing planning technique that uses abduction can be combined with Inductive Logic Programming to use and adapt the knowledge of the agent. The use of an interactive inductive logic programming method allows the system to learn from experiments.