Data Acquisition for Argument Search

The args.me Corpus

verfasst von

Yamen Ajjour, Henning Wachsmuth, Johannes Kiesel, Martin Potthast, Matthias Hagen, Benno Stein

Abstract

Argument search is the study of search engine technology that can retrieve arguments for potentially controversial topics or claims upon user request. The design of an argument search engine is tied to its underlying argument acquisition paradigm. More specifically, the employed paradigm controls the trade-off between retrieval precision and recall and thus determines basic search characteristics: Compiling an exhaustive argument corpus offline benefits precision at the expense of recall, whereas retrieving arguments from the web on-the-fly benefits recall at the expense of precision. This paper presents the new corpus of our argument search engine args.me, which follows the former paradigm. We freely provide the corpus to the community. With 387 606 arguments it is one of the largest argument resources available so far. In a qualitative analysis, we compare the args.me corpus acquisition paradigm to that of two other argument search engines, and we report first empirical insights into how people search with args.me.

Externe Organisation(en)

Bauhaus-Universität Weimar
Universität Paderborn
Universität Leipzig
Martin-Luther-Universität Halle-Wittenberg

Typ

Aufsatz in Konferenzband

Seiten

48-59

Anzahl der Seiten

12

Publikationsdatum

01.09.2019

Publikationsstatus

Veröffentlicht

Peer-reviewed

Ja

ASJC Scopus Sachgebiete

Theoretische Informatik, Informatik (insg.)

Elektronische Version(en)

https://doi.org/10.1007/978-3-030-30179-8_4 (Zugang: Geschlossen)