Contextual Language Models For Ranking Answers To Natural Language Definition Questions
Computational Intelligence
TextWiki: a superlative resource
Language Resources and Evaluation
Hi-index | 0.00 |
In this paper, we present an approach to answering "Other" questions using the notion of interest marking terms. "Other" questions have been introduced in the TREC-QA track to retrieve other interesting factsabout a topic. To answer these types of questions, our system extracts from Wikipedia articles a list of interest-markingterms related to the topic and uses them to extract and score sentences from the document collection where the answer should be found. Sentences are then re-ranked using universal interest-markers that are not specific to the topic. The top sentences are then returned as possible answers. When using the 2004 TREC data for development and 2005 data for testing, the approach achieved an F-score of 0.265, placing it among the top systems.