Spoken Question Answering
Sophie Rosset, Olivier Galibert, Lori F Lamel · 2011
This chapter covers Question-Answering (QA) from spoken documents (referred to as QAst), but also beyond, where questions are also spoken. It presents some specific aspects that must be considered when handling speech data for question answering. The chapter also presents the main evaluation campaigns that have been carried out in the question-answering domain, the majority of which have addressed only written language. To the best of our knowledge, to date there has only been one evaluation for spoken QA. However, since the general problematics are the same, the author believes it important to also present the larger view including written QA. The chapter provides the important considerations of QA evaluations, followed by a detailed presentation of the QAst campaigns. It describes and compares different approaches and systems for QA, with a focus on approaches used for spoken language. The chapter reviews recent projects or work addressing spoken QA. Controlled Vocabulary Terms learning (artificial intelligence); speech processing