Dokument: Bridging the Gap between Online Discussions and Formal Models of Argumentation
| Titel: | Bridging the Gap between Online Discussions and Formal Models of Argumentation | |||||||
| URL für Lesezeichen: | https://docserv.uni-duesseldorf.de/servlets/DocumentServlet?id=50163 | |||||||
| URN (NBN): | urn:nbn:de:hbz:061-20190710-094214-3 | |||||||
| Kollektion: | Dissertationen | |||||||
| Sprache: | Englisch | |||||||
| Dokumententyp: | Wissenschaftliche Abschlussarbeiten » Dissertation | |||||||
| Medientyp: | Text | |||||||
| Autor: | Neugebauer, Daniel [Autor] | |||||||
| Dateien: |
| |||||||
| Beitragende: | Prof. Dr. Rothe, Jörg [Gutachter] [im Online-Personal- und -Vorlesungsverzeichnis LSF anzeigen] Jun.-Prof. Dr. Baumeister, Dorothea [Gutachter] [im Online-Personal- und -Vorlesungsverzeichnis LSF anzeigen] | |||||||
| Dewey Dezimal-Klassifikation: | 000 Informatik, Informationswissenschaft, allgemeine Werke » 004 Datenverarbeitung; Informatik | |||||||
| Beschreibung: | Formal models of argumentation in the field of artificial intelligence (AI) research were originally developed as a tool for logical inference on the basis of incomplete or inconsistent knowledge. Recently, AI argumentation formalisms are also being used to represent and evaluate real-world argumentations. Software tools that support argumentation can benefit from the semantic evaluation mechanisms that are provided by formal argumentation models, while the application to real-world scenarios improves the scope and relevance of formal models of argumentation in AI.
This research is part of the Ph.D. program "Online Participation," supported by the North Rhine-Westphalian funding scheme "Fortschrittskollegs." One of the program's goals is to improve the quality of software-supported argumentation. This thesis contributes to this improvement by strengthening the connection between formal argumentation models from AI and software-supported discussions among humans on both sides. The first part of the contribution of this thesis is an extension of the formal model of abstract argumentation frameworks. In its basic form, an argumentation framework is given by a set of arguments and a binary attack relation on these arguments. In the proposed extended model of incomplete argumentation frameworks, individual arguments and individual attacks may be uncertain—such elements may or may not be part of the discussion. The extended model allows the representation of a wider range of scenarios that can arise in real argumentation settings, such as intermediate states in elicitation processes or the restricted knowledge of a single participant about the state of a discussion. The technical contribution of this part of the thesis is a full analysis of how the added notion of uncertainty affects the computational complexity of core reasoning tasks in abstract argumentation—namely, the verification and acceptance problems—for various evaluation semantics. Compared to the computational complexity of the respective problems for standard argumentation frameworks, for incomplete argumentation frameworks we observe a jump in complexity for most variants of the acceptance problems, but not for most variants of the verification problem. The second part of the contribution of this thesis is the development and implementation of translations from discussion data generated by the D-BAS web tool for dialog-based argumentation to three different AI argumentation formalisms, namely, abstract argumentation frameworks, abstract dialectical frameworks, and the ASPIC+ framework. The translations are proved to satisfy established quality criteria. We developed the tool DABASCO to implement these translations, which enables D-BAS to utilize the full range of existing software tools for reasoning problems in the three argumentation models. The argument pipeline consisting of D-BAS, DABASCO, and these tools can automatically determine whether, e.g., a statement is acceptable or whether a participant's opinion is consistent, thus improving feedback for operators and participants of discussions. | |||||||
| Lizenz: |  Urheberrechtsschutz | |||||||
| Fachbereich / Einrichtung: | Mathematisch- Naturwissenschaftliche Fakultät » WE Informatik » Computational Complexity Theory/Cryptology | |||||||
| Dokument erstellt am: | 10.07.2019 | |||||||
| Dateien geändert am: | 10.07.2019 | |||||||
| Promotionsantrag am: | 19.03.2019 | |||||||
| Datum der Promotion: | 02.07.2019 |
