Resolving Entity Morphs in Censored Data

Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics, pages 1083–1093, Sofia, Bulgaria, August 4-9 2013. c⃝2013 Association for Computational Linguistics Resolving Entity Morphs in Censored Data Hongzhao Huang1, Zhen Wen2, Dian Yu1, Heng Ji1, Yizhou Sun3, Jiawei Han4, He Li5 1Computer Science Department and Linguistics Department, Queens College and Graduate Center, City University of New York, New York, NY, USA 2IBM T. J. Watson Research Center, Hawthorne, NY, USA 3College of Computer and Information Science, Northeastern University, Boston, MA, USA 4Computer Science Department, Univerisity of Illinois at Urbana-Champaign, Urbana, IL, USA 5Admaster Inc., China {hongzhaohuang1,yudiandoris1,hengjicuny1, liheact5}@gmail.com, zhenwen@us.ibm.com2, yzsun@ccs.neu.edu3, hanj@illinois.edu4 Abstract In some societies, internet users have to create information morphs (e.g. “Peace West King” to refer to “Bo Xilai”) to avoid active censorship or achieve other com- munication goals. In this paper we aim to solve a new problem of resolving en- tity morphs to their real targets. We ex- ploit temporal constraints to collect cross- source comparable corpora relevant to any given morph query and identify target can- didates. Then we propose various novel similarity measurements including surface features, meta-path based semantic fea- tures and social correlation features and combine them in a learning-to-rank frame- work. Experimental results on Chinese Sina Weibo data demonstrate that our ap- proach is promising and significantly out- performs baseline methods1. 1 Introduction Language constantly evolves to maximize com- municative success and expressive power in daily social interactions. The proliferation of online so- cial media significantly expedites this evolution, as new phrases triggered by social events may be disseminated rapidly in social media. To automati- cally analyze such fast evolving language in social media, new computational models are demanded. In this paper, we focus on one particular lan- guage evolution that creates new ways to commu- nicate sensitive subjects because of the existence of internet information censorship. We call this 1Some of the resources and open source programs devel- oped in this work are made freely available for research pur- pose at http://nlp.cs.qc.cuny.edu/Morphing.tar.gz phenomenon information morph. For example, when Chinese online users talk about the former politician “Bo Xilai”, they use a morph “Peace West King” instead, a historical figure four hun- dreds years ago who governed the same region as Bo. Morph can be considered as a special case of alias used for hiding true entities in ma- licious environment (Hsiung et al., 2005; Pantel, 2006). However, social network plays an impor- tant role in generating morphs. Usually morphs are generated by harvesting the collective wisdom of the crowd to achieve certain communication goals. Aside from the purpose of avoiding cen- sorship, other motivations for using morph include expressing sarcasm/irony, positive/negative senti- ment or making descriptions more vivid toward some entities or events. Table 1 presents the wide range of cases that are used to create the morphs. We can see that a morph can be either a regular term with new meaning or a newly created term. Morph Target Motivation Peace West King Bo Xilai Sensitive Blind Man …

Content truncated to 500 words