This page contains a summary of my main research interests.
My main research area is Case-Based Reasoning (CBR). Since my master’s thesis, I have been working on supporting and automating knowledge acquisition tasks when building CBR systems. CBR is a methodology that utilizes previous knowledge and experiences to solve future problems. The CBR methodology is centered around cases that describe situations as problem-solution pairs. The methodology describes how a CBR system finds a solution for a given problem description: first, it searches most similar existing cases and then adapts them, e.g., through heuristics. Next, the newly created solution is tested before the system can learn the problem-solution pair. In my research, I am interested in how to build CBR systems in various application domains and focus on the investigation of determining the relevant features, learning similarity measures from data, develop retrieval strategies to find relevant cases, and increase the variety of returned results. Besides addressing single parts of the CBR methodology, I also work on an overarching process that describes how domain expert knowledge can be utilized in CBR systems.
As an active member of the CBR community, I have co-chaired the International Conference on Case-Based Reasong (ICCBR) in 2019, hosted it in Trondheim in 2017, and have been organizing workshops and side-events since 2008.
Artificial Intelligence Applications
A pivotal aspect of my work is the collaboration with domain experts to develop AI applications. Many AI applications focus on secondary care and healthcare professionals, while primary care has received less focus. In our collaborations, we mainly address problems in primary care that affect many people and we develop digital tools to improve their diagnosis and treatments. The challenges we address are how to develop AI tools that are understood and trusted by domain experts. We use randomized controlled trials (RCT) as the gold standard to test their effectiveness, which is extremely rare in AI research. This work has been carried out in the selfBACK and BACK-UP EU projects. Current projects addressing treatment planning and return-to-work have recently started: SupportPrim and SmaRTWork.
Open Source Software
I have been working on the open-source tool myCBR since 2010 and have collaborated with researchers from the CBR community to provide tools that facilitate the application of CBR. The myCBR tool is part of the CBR education at NTNU and other universities. The work of my Ph.D. students and postdoctoral fellows contributed to extensions of the tool. The results of our work is made available here: https://github.com/orgs/ntnu-ai-lab/teams/mycbr/repositories.