Qme! : A Speech-based Question-Answering system on Mobile Devices

Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the ACL, pages 55–63, Los Angeles, California, June 2010. c⃝2010 Association for Computational Linguistics Qme! : A Speech-based Question-Answering system on Mobile Devices Taniya Mishra AT&T Labs-Research 180 Park Ave Florham Park, NJ taniya@research.att.com Srinivas Bangalore AT&T Labs-Research 180 Park Ave Florham Park, NJ srini@research.att.com Abstract Mobile devices are becoming the dominant mode of information access despite being cumbersome to input text using small key- boards and browsing web pages on small screens. We present Qme!, a speech-based question-answering system that allows for spoken queries and retrieves answers to the questions instead of web pages. We present bootstrap methods to distinguish dynamic questions from static questions and we show the benefits of tight coupling of speech recog- nition and retrieval components of the system. 1 Introduction Access to information has moved from desktop and laptop computers in office and home environments to be an any place, any time activity due to mo- bile devices. Although mobile devices have small keyboards that make typing text input cumbersome compared to conventional desktop and laptops, the ability to access unlimited amount of information, almost everywhere, through the Internet, using these devices have made them pervasive. Even so, information access using text input on mobile devices with small screens and soft/small keyboards is tedious and unnatural. In addition, by the mobile nature of these devices, users often like to use them in hands-busy environments, ruling out the possibility of typing text. We address this issue by allowing the user to query an information repos- itory using speech. We expect that spoken language queries to be a more natural and less cumbersome way of information access using mobile devices. A second issue we address is related to directly and precisely answering the user’s query beyond serving web pages. This is in contrast to the current approach where a user types in a query using key- words to a search engine, browses the returned re- sults on the small screen to select a potentially rele- vant document, suitably magnifies the screen to view the document and searches for the answer to her question in the document. By providing a method for the user to pose her query in natural language and presenting the relevant answer(s) to her question, we expect the user’s information need to be fulfilled in a shorter period of time. We present a speech-driven question answering system, Qme!, as a solution toward addressing these two issues. The system provides a natural input modality – spoken language input – for the users to pose their information need and presents a col- lection of answers that potentially address the infor- mation need directly. For a subclass of questions that we term static questions, the system retrieves the answers from an archive of human generated an- swers to questions. This ensures higher accuracy for the answers retrieved (if found in the archive) and also allows us to retrieve related questions on the user’s …

Content truncated to 500 words