Please use this identifier to cite or link to this item:
|Title:||PolyUCOMP in TAC 2011 entity linking and slot filling|
|Publisher:||National Institute of Standards and Technology|
|Source:||Proceedings of the Fourth Text Analysis Conference (TAC 2011), November 14-15, 2011, National Institute of Standards and Technology, Gaithersburg, Maryland, USA.|
|Abstract:||The PolyUCOMP team participated in two TAC-KBP2011 tasks: Regular Entity Linking and Regular Slot Filling.|
For the entity linking task, a three-step entity linking system is developed. Similar to some systems in KBP 2010, a list of possible candidates are first selected. Then the best candidate is identified to decide whether a link exists. In addition, a document clustering algorithm is used to group NIL queries. This system incorporates Wikipedia linking information and other textual and contextual features. Our system produces a high answer coverage and accurate linking result. However, the NIL detection system brings significant loss in the final F-score.
For the slot filling task, we developed a system which combines query expansion and pattern-based reasoning. In expanding person queries, name variants are produced using different rules. For organization acronym queries, an abbreviation extraction technique is employed. Manually collected triggers are used for extracting other slot values. Our system ranked above median among all the participating systems.
|Description:||The Text Analysis Conference (TAC) is organized by the U.S. National Institute of Standards and Technology (NIST).|
|Rights:||Reproduced with permission of the author.|
|Appears in Collections:||COMP Conference Papers & Presentations|
Files in This Item:
|Xu_PolyUCOMP_TAC_2011.pdf||392.99 kB||Adobe PDF||View/Open|
All items in the PolyU Institutional Repository are protected by copyright, with all rights reserved, unless otherwise indicated. No item in the PolyU IR may be reproduced for commercial or resale purposes.