معرفی
Kary Främling is a Professor in Data Science with emphasis on data analysis and machine learning at the Department of Computing Science, Umeå University. He serves as the Head of the Explainable AI (XAI) team and holds the prestigious position of WASP Professor within the Wallenberg AI, Autonomous Systems and Software Program, focusing on making AI systems transparent and understandable to end-users in real-world contexts.
Professor Främling's research centers on Explainable Artificial Intelligence (XAI), particularly "outcome explanation" - explaining and justifying AI system results. His foundational work includes the Contextual Importance and Utility (CIU) method developed during his 1991-1996 PhD. His research spans Intelligent products, Internet of Things (IoT), Digital Twin technology, and Systems of Systems, emphasizing that explainability is crucial for ensuring AI remains "humane" - able to communicate its reasoning appropriately for different users and situations.
Analysis of his recent publications reveals a strong focus on practical XAI implementations across healthcare, smart buildings, IoT systems, and image classification. His work bridges theoretical AI research with practical tool development (py-ciu, ciu R packages) and application-specific implementations, consistently emphasizing contextually appropriate explanations tailored to specific users and scenarios.
His scientific recognition includes:
- WASP Professor (Wallenberg AI, Autonomous Systems and Software Program)
As Head of the XAI team, Professor Främling leads research developing and applying explainability methods to real-world AI systems. He teaches "Data preprocessing and visualisation (5DV217)" and has been featured in ACM Computing Surveys, Springer LNCS, and various IEEE/MDPI publications. His work involves international collaborations across Europe, focusing on making black-box AI models transparent and trustworthy through context-aware explanations.
Professor Främling leads the XAI research group at Umeå University, developing and applying the Contextual Importance and Utility method across domains including healthcare diagnostics, smart building energy management, agricultural technology, and affect detection systems. The group collaborates with industry partners to implement explainable AI in practical systems while advancing theoretical foundations of explanation generation.
