Multimedia answering: Enriching text QA with media information

Abstract

Existing community question-answering forums usually provide only textual answers. However, for many questions, pure texts cannot provide intuitive information, while image or video contents are more appropriate. In this paper, we introduce a scheme that is able to enrich text answers with image and video information. Our scheme investigates a rich set of techniques including question/answer classification, query generation, image and video search reranking, etc. Given a question and the community-contributed answer, our approach is able to determine which type of media information should be added, and then automatically collects data from Internet to enrich the textual answer. Different from some efforts that attempt to directly answer questions with image and video data, our approach is built based on the community-contributed textual answers and thus it is more feasible and able to deal with more complex questions. We have conducted empirical study on more than 3,000 QA pairs and the results demonstrate the effectiveness of our approach.