| Overview of the TREC 2001 question answering track (2001) | |||||||||||||||
Abstract | |||||||||||||||
| The TREC question answering track is an effort to bring the benefits of loxge-scale evaluation to beox on the question answering problem. In its third yeox, the track continued to focus on retrieving small snippets of text that contain an answer to a question. However, several new conditions were added to increase the realism, nd the difficulty, of the task. In the main task, questions were no longer guoxanteed to have n nswer in the collection; systems returned a response of 'NIL ' to indicate their belief that no nswer was present. In the new list task, systems assembled a set of instances as the response for a question, requiring the ability to distinguish aong instances found in multiple documents. Another new task, the context task, required systems to track discourse objects through a series of questions. The TREC 2001 question answering (QA) track was the third running of a QA track in TREC. The goal of the track has remained the same each year: to foster research on systems that retrieve answers rather than documents in response to a question, with a particular emphasis on systems that can function in unrestricted domains. Systems are given a large corpus of newspaper and newswire articles and a set of closed-class questions such as Who invented the paper clip?. They return a short ( _ 50 bytes) text snippet and a document as a response to a question, where the snippet contains an answer to the question and the | |||||||||||||||
Publication details | |||||||||||||||
| |||||||||||||||