Automatically Analyzing Scientific Argumentation and Reasoning

Affiliation: TU Darmstadt

Associated Since: January 2016

• Dr. Johannes Daxenberger (UKP Lab, Fachbereich Informatik, TU Darmstadt)
• Prof. Dr. Iryna Gurevych (UKP Lab, Fachbereich Informatik, TU Darmstadt)

CEDIFOR-Partners:

• Prof. Dr. Frank Fischer (Empirische Pädagogik und Pädagogische Psychologie, LMU München)

This project aims to analyze scientific argumentation and reasoning in professional practice. To this end, we carried out an automatic inspection on existing datasets from the REASON Program at the LMU Munich. Among other things, we compared the reasoning chains in think-aloud protocols by teacher students and probation officers. Ultimately, we aim to detect similar or divergent patterns in the argumentative structures of the datasets from the professional practices.This project involves researchers from educational and psychological domains, and at the same time wants to improve methods automatic text classification.

A system which demonstrates the automatic classification of epistemic types and resulting performance can be found here.

Publications:

• Daxenberger, J., Csanadi, A., Ghanem, C., Kollar, I., & Gurevych, I. (2018). Domain-Specific Aspects of Scientific Reasoning and Argumentation: Insights from Automatic Coding. In F. Fischer, C. Chinn, K. Engelmann, & J. Osborne (Eds.), Scientific Reasoning and Argumentation: The Roles of Domain-Specific and Domain-General Knowledge (p. to appear). Taylor & Francis.
• Tobias Kahse (2017). Multi-Task Learning for Argumentation Mining. Master Thesis: Technische Universität Darmstadt.
• Csanadi, A., Daxenberger, J., Ghanem, C., Kollar, I., Fischer, F., & Gurevych, I. (2016). Automated Text Classification to Capture Scientific Reasoning and Argumentation Processes in Different Professional Problem Solving Contexts. Extended Abstract at Twenty-sixth Annual Meeting of the Society for Text & Discourse.
• Lerner, P., Csanadi, A., Daxenberger, J., Flekova, L., Ghanem, C., Kollar, I., Fischer, F., Gurevych, I. (2016). A User Interface for the Exploration of Manually and Automatically Coded Scientific Reasoning and Argumentation. In Proceedings of the 12th International Conference of the Learning Sciences.

Vorträge:

• Invited Talk at the International Workshop “Interplay of Domain-Specific and Domain-General Aspects of Scientific Reasoning and Argumentation Skills” at the Center for Advanced Studies (Ludwig-Maximilians-Universität München): 6-8 July 2016