Invited Talk - Katarzyna Budzynska

Title: Computational Ethos (Katarzyna Budzynska, Laboratory of The New Ethos, Warsaw University of Technology, Poland)

Bio: Katarzyna Budzynska studies argumentation and persuasion linking philosophy, rhetoric and natural language processing. She heads the Laboratory of The New Ethos at the Warsaw University of Technology in Poland. The group develops models and techniques to process the use of ethos, the character of the speaker, in natural language in order to predict the results of presidential elections and detect trolls and cyber bullies in social media. She has published 2 books and over 80 peer-reviewed papers in journals such as "Synthese", "ACM Transactions on Internet Technology" and "Artificial Intelligence". Her research attracted 1m EUR grant funding in Polish, Swiss, British, German and European programs including the recent project on Interactive Natural Language Technology for Explainable Artificial Intelligence, NL4XAI, https://nl4xai.eu/ (MSCA-ITN, H2020).

Abstract: Ethos, the character of the speaker, is a powerful tool used to influence others through communication. Together with logos (argumentation) and pathos (emotions of the audience), they constitute the key elements of Aristotle's Rhetoric. This talk introduces Computational Ethos, a model of ethotic strategies, derived from and verified on real-life argument data, which is applied to several domains in order to implement technologies such as ethos mining and ethos analytics. In the NL4XAI project, Interactive Natural Language Technology for Explainable Artificial Intelligence, foundations laid most recently in (Duthie, Budzynska, 2018; Budzynska, Koszowy, Pereira-Fari{\~n}a, forthcoming) will be applied to create customised interactive argumentation schemes for XAI. The objective is to define, design, develop and validate a new methodology for argumentation schemes tailored for planning and handling the discourse history in interactive conversational agents aimed at explaining AI in natural language. The customisation will be achieved through the creation of a profile of the user and then adaptation of argumentation to match this profile.

Accepted Papers

• Joanna Misztal-Radecka (AGH University of Science and Technology, Ringier Axel Springer); Bipin Indurhkya (Jagiellonian University. "Persona Prototypes for Improving the Qualitative Evaluation of Recommendation Systems"
  


• Cataldo Musto (University of Bari); Fedelucio Narducci (Polytechnic University of Bari); Marco Polignano (University of Bari); Marco de Gemmis (University of Bari); Pasquale Lops (University of Bari); Giovanni Semeraro (University of Bari). "Towards Queryable User Profiles: Introducing Conversational Agents in a Platform for Holistic User Modeling"
  


• Diana C Hernandez-Bocanegra (University of Duisburg); Jürgen Ziegler (University of Duisburg); Tim Donkers (University of Duisburg). "Effects of argumentative explanation types on the perception of review-based recommendations"
  


• Sidra Naveed (University of Duisburg); Benedikt Loepp (University of Duisburg); Jürgen Ziegler (University of Duisburg). "On the Use of Feature-based Collaborative Explanations: An Empirical Comparison of Explanation Styles"
  


• Mirko Polato (University of Padova); Tommaso Carraro (University of Padova); Fabio Aiolli (University of Padova). "A look inside the black-box: towards the interpretability of Conditioned Variational Autoencoder for Collaborative Filtering"
  


• Stefan Hirschmeier (University of Cologne); Detlef Schoder (University of Cologne). An Approach to Explanations for Public Radio Recommendations