Proceedings of the Second International Conference Computational Linguistics in Bulgaria

Svetla Koeva​, Karel Oliva​, Mariana Damova​, Mila Dimitrova­vulchanova​ Norwegian, Preslav Nakov, Radev
2016 Proceedings of the Second International Conference Computational Linguistics in Bulgaria   unpublished
PREFACE We are excited to welcome you to the second edition of the International Conference ​ Computational Linguistics in Bulgaria (CLIB 2016) in Sofia, Bulgaria! CLIB aspires to foster the NLP community in Bulgaria and further the cooperation among researchers working in NLP for Bulgarian around the world. The need for a conference dedicated to NLP research dealing with or applicable to Bulgarian has been felt for quite some time. We believe that building a strong community of researchers and
more » ... teams who have chosen to work on Bulgarian is a key factor to meeting the challenges and requirements posed to computational linguistics and NLP in Bulgaria. We share the hope that CLIB will establish itself as an international forum for sharing high­quality scientific work in all areas of computational linguistics and NLP and will grow in scope and scale with each new edition. The CLIB community will be dedicated to supporting the creation and improvement of advanced NLP resources, tools and technologies for mono­ and multilingual language processing, machine translation and translation aids, content creation, localisation and personalisation, speech recognition and generation, information retrieval and information extraction. The Conference was made possible due to the hard work of many people. We would like to thank the authors who trusted us and submitted their contributions to CLIB 2016. Their efforts and high­quality research are the chief factor that enabled us to create an interesting and solid scientific programme. We would also like to thank our industrial participants for sharing their insights, ideas and know­how with the research community. We would like to express our sincere gratitude to the members of the Programme Committee, who accepted to join us and invested a lot of expertise to provide valuable feedback to the authors. Special thanks are due to Prof. Svetla Koeva, who is the person behind the whole CLIB concept. We hope that CLIB 2016 will be a useful and productive experience that we all will enjoy! CLIB 2016 Organising Committee iv PROGRAMME COMMITTEE Exposing Paid Opinion Manipulation Trolls in News Community Forums The practice of using opinion manipulation trolls has been reality since the rise of Internet and community forums. It has been shown that user opinions about products, companies and politics can be influenced by posts by other users in online forums and social networks. This makes it easy for companies and political parties to gain popularity by paying for "reputation management" to people or companies that write in discussion forums and social networks fake opinions from fake profiles. During the 2013­2014 Bulgarian protests against the Oresharski cabinet, social networks and news community forums became the main "battle grounds" between supporters and opponents of the government. In that period, there was a very notable presence and activity of government supporters in Web forums. In series of leaked documents in the independent Bulgarian media Bivol, it was alleged that the ruling Socialist party was paying Internet trolls with EU Parliament money. Allegedly, these trolls were hired by a PR agency and were given specific instructions what to write. A natural question is whether such trolls can be found and exposed automatically. This is a very hard task, as there is no enough data to train a classifier; yet, it is possible to obtain some test data, as these trolls are sometimes caught and widely exposed (e.g., by Bivol). Yet, one still needs training data. We solve the problem by assuming that a user who is called a troll by several different people is likely to be one, and one who has never been called a troll is unlikely to be such. We compare the profiles of (i) paid trolls vs. (ii) "mentioned" trolls vs. (iii) non­trolls, and we further show that a classifier trained to distinguish (ii) from (iii) does quite well also at telling apart (i) from (iii). KEYNOTE TALK