2012 AFrameworkforRobustDiscoveryofE

From GM-RKB
Jump to navigation Jump to search

Subject Headings:

Notes

Cited By

Quotes

Author Keywords

Abstract

Entity synonyms are critical for many applications like information retrieval and named entity recognition in documents. The current trend is to automatically discover entity synonyms using statistical techniques on web data. Prior techniques suffer from several limitations like click log sparsity and inability to distinguish between entities of different concept classes. In this paper, we propose a general framework for robustly discovering entity synonym with two novel similarity functions that overcome the limitations of prior techniques. We develop efficient and scalable techniques leveraging the MapReduce framework to discover synonyms at large scale. To handle long entity names with extraneous tokens, we propose techniques to effectively map long entity names to short queries in query log. Our experiments on real data from different entity domains demonstrate the superior quality of our synonyms as well as the efficiency of our algorithms. The entity synonyms produced by our system is in production in Bing Shopping and Video search, with experiments showing the significance it brings in improving search experience.

References

;

 AuthorvolumeDate ValuetitletypejournaltitleUrldoinoteyear
2012 AFrameworkforRobustDiscoveryofEDong Xin
Surajit Chaudhuri
Tao Cheng
Kaushik Chakrabarti
A Framework for Robust Discovery of Entity Synonyms10.1145/2339530.23397432012