Dynamically Improving Explanations: A Revision-Based Approach to Explanation Generation
Recent years have witnessed rapid progress in explanation generation. Despite these ad vances, the quality of prose produced by ex planation generators warrants significant im provement. Revision-based explanation gener ation offers a promising means for improving explanations at runtime. In contrast to single- draft explanation generation architectures, a revision-based generator could dynamically cre ate, evaluate, and refine multiple drafts of explanations. However, because of the in herent complexity of revision, previous multi- sentential revision-based approaches have not scaled up. We have developed a scalable revision-based model of explanation generation that dynamically improves multi-sentential ex planations. By operating on abstract discourse plans encoded in a minimalist representation, it combats both the conceptual complexities and the efficiency problems posed by revision. This approach has been implemented in REVISOR, a unification-based revision system. Evaluations of REVISOR'S performance in generating a cor pus of extended multi-sentential scientific ex planations yielded encouraging results.