Toward a framework for assessing explanation quality (Project INF)
Complementarily to subprojects across research areas A, B, and C, the members of Project INF investigate how the quality and success of an explanation can be measured. For this purpose, the computer scientists will first analyze natural language explanations, such as written records from other projects or explanatory dialogues from internet forums. Based on the findings from this analysis, they will then develop tools and criteria that can be applied to explanations. In addition, the project members will provide TRR researchers with a glossary of key concepts. The aim of this project is to establish an understanding of what makes an explanation successful. With these findings, the researchers will contribute to a comprehensive theory of the co-construction of explanations as elaborated in the TRR.
Publications
Wachsmuth, H., & Alshomary, M. (2022). "Mama Always Had a Way of Explaining Things So I Could Understand'': A Dialogue Corpus for Learning to Construct Explanations. Proceedings of the 29th International Conference on Computational Linguistics arXiv. https://doi.org/10.48550/ARXIV.2209.02508