METHOD FOR CREATING PREDICTIVE KNOWLEDGE STRUCTURES FROM EXPERIENCE IN AN ARTIFICIAL AGENT

摘要

A method comprises building a forecast for an autonomous agent. The building at least comprises assigning a selected parameter of the autonomous agent to a state value, adding a new policy to a set of policies where the new policy maps actions of the autonomous agent for optimizing the state value, and adding a new forecast to a set of forecasts where the forecast at least comprises a prediction of a next state of the autonomous agent following execution of the new policy. The mapped actions are initiated according to the new policy. A state of the autonomous agent is evaluated following completion of the mapped actions. The evaluation at least comprises comparing the state with the forecast. Whether to build an additional forecast is determined. The determining is at least in part based on the evaluation.