Entering a new world: the minimal amount of knowledge to act as a trustworthy adviser using case-based explanations in a new domain

Jakob Michael Schönborn

In: Mirjam Minor (Hrsg.). Doctoral Consortium. ICCBR Doctoral Consortium (ICCBR-DC-2018), located at International Conference on Case-Based Reasoning, July 9-12, Stockholm, Sweden, Pages 224-228, ICCBR, 8/2018.


During the last years, there are multiple approaches to adapt Case-Based Reasoning to other domains than usually used before. Nevertheless, starting to develop a working Case-Based Reasoning system with the known issues of how to initialize a well-structured knowledge base and especially how to gather the required knowledge seems to be an issue. On top, the user's acceptance of decisions made by artificial intelligence agents is more skeptical than welcoming. Therefore, plausible explanations have to be generated for each decision made so that the user can develop trust in these. The problem is to determine how much knowledge in the given domain is actually needed to act as a trustworthy adviser and in general how to structure the explanation so that it will be accepted by the user. When building up a new explanation-aware CBR system, the process itself of creating this system should be capable of explaining itself. on top, the resulting CBR system should also be able to offer explanations.

Weitere Links

2018-iccbr-dc-jakob_ws_proceedings.pdf (pdf, 1 MB )

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence