BAbI: A Test of Commonsense Ability

The BAbI benchmark presents a difficult set of tasks designed to evaluate the skills of AI systems in interpreting commonsense knowledge. It comprises a wide range of situations that require logic about everyday concepts. By evaluating how well AI models can solve these problems, researchers aim to better understand the essence of commonsense reaso

read more