People in all societies argue, not only to persuade others of their opinions, but also to resolve disputes, to obtain new insights, or to make decisions. Arguments play a major role for most controversial topics in everyday life and politics, such as how to balance work and private life, whether feminism is still needed, or how to react on Trump's import tariffs. As as result, computational argumentation — i.e., the empirical, usually data-driven analysis and synthesis of natural langugage argumentation — has recently been seeing an influx of research activities, in academia as well as in industry, with applications such as argument search engines and intelligent personal assistants. Arising research questions in the context of information retrieval include, for example, how to reliably crawl arguments on the web, how do users query for arguments, or, what arguments are considered most relevant for a given topic.
This tutorial will introduce main concepts and methods from computational argumentation, with an emphasis on retrieval and analysis. Building upon foundations from argumentation theory, linguistics, and rhetoric, we will cover the mining of arguments from natural language text, the assessment of an argument's stance and quality, the interpretation of argument graph structures, and more. The contents will be presented partly in overview form, partly more in depth, accompanied by demos and hands-on tasks. On this basis, the role of computational argumentation in information retrieval research will be explored in discussion with the participants.
The main objective of the tutorial is to enable information retrieval researchers to consider computational argumentation for their studies as well as to convey the basics of integrating it into their own work.