Understanding Text Using Agent Based Models

The paper proposes a novel approach to text understanding and generation focusing on short and simple stories. The proposed approach attempts to understand and generate stories by creating an explainable, agent-based world model of the story. In the model, agents have clearly defined goals, based on which they automatically evaluate their next actions. We demonstrate our approach on the story of ‘Little Red Riding Hood’, simulating it as a sequence of almost 50 actions.