Profiling humans, profiling bots, profiling you.

Author profiling studies how language is shared by people. This helps in identifying aspects such as gender, age, native language, or even personality. Author profiling is a problem of growing importance in forensics, security, and marketing. E.g., from a forensic linguistics perspective one would like being able to know the profile of the author of a harassing text message; from a marketing viewpoint, companies may be interested in knowing, on the basis of the analysis of blogs and online product reviews, the demographics of people that like or dislike their products. Our  focus will be on author profiling in social media since we are mainly interested in everyday language.

After having addressed since 2013 at the PAN Lab at CLEF (https://pan.webis.de/) several aspects of author  profiling in social media (age and gender, personality, language variety, and gender from a multimodal perspective), in 2019 we investigated the feasibility of distinguishing whether the author of a Twitter feed is a bot or a human. In fact, social media bots may influence users with commercial, political or ideological purposes. For example, bots could artificially inflate the popularity of a product by promoting it and/or writing positive ratings, as well as undermine the reputation of competitive products through negative valuations. The threat is even greater when the purpose is political such as for the Brexit referendum or the US Presidential election (fearing the effect of this influence, the German political parties have rejected the use of bots in their electoral campaign for the general elections). Furthermore, bots are commonly related to fake news spreading. Therefore, to approach the identification of bots from an author profiling perspective is of high importance. In this talk we will briefly describe some of the these years' PAN author profiling shared tasks.

About the Speaker:  Paolo Rosso (http://users.dsic.upv.es/~prosso/) is full professor at the Universitat Politecnica de Valencia, Spain where he is also member of the PRHLT research center. His research interests focus mainly on author profiling, irony detection, opinion spam detection, and plagiarism detection. Since 2009 he has been involved in the organisation of PAN benchmark activities at CLEF and at FIRE evaluation forums, mainly on plagiarism / text reuse detection and author profiling. At SemEval he has been co-organiser of shared tasks on sentiment analysis of figurative language in Twitter (2015), and on multilingual detection of hate speech against immigrants and women in Twitter (2019). He has been PI of national and also international research projects funded by EC and U.S. Army Research Office. Recently, in collaboration with Carnegie Mellon University, he is involved in a project funded by Qatar National Research Fund on author profiling for cyber-security, which aims to profile who is behind to threat messages. He serves as deputy steering committee chair for the CLEF conference and as associate editor for the Information Processing & Management journal. He has been chair of *SEM-2015, and organization chair of CERI-2012, CLEF-2013 and EACL-2017. He is the author of 200+ papers, published in journals, book chapters,  conference and workshop proceedings.