Competence Based Question Generation

AbstractModels of natural language understanding often rely on question answering and logical inference benchmark challenges to evaluate the performance of a system. While informative, such task-oriented evaluations do not assess …...