RESEARCH BIO
Senior lecturer (Docent, PhD) in computer science, with a research background in analysis of complex systems, distributed systems and databases, and applied ML.
The research focus in recent years includes explainable and trustworthy AI systems. Along with dramatically increased capabilities of algorithms, there is a now need for designing AI system intentionally for capabilities of for instance explaining AI model uncertainty and AI model assumptions. The adaptation degree of AI used in new critical application areas will depend on how transparent automated decisions will be.
Since 2016, Dr. Mathiason has developed multiple collaborations with the Swedish steel industry with the intention to increase utilization of AI in the steel industry. Leading the recent projects Dataflow (Outokumpu, Jernkontoret, funded by VINNOVA), Swedish Metal (SSAB and Sandvik SMT, funded by KK), and now INSITE-X (Ovako, Ouokumpu, funded by VINNOVA) several novel papers were published on advanced process analysis using deep learning (DL): In Dataflow, DL was used for revealing the exact part steel sheet shapes determining telescoping errors, while in Swedish Metal a complex steel converter process was modelled using DL. The new project INSITE-X (Ovako and Outokumpu) will use DL to learn a detailed digital twin from only high-resolution data to model dynamic in-production machine behavior.
In 2019, Dr. Mathiason was appointed by Jernkontoret for delivering texts on the strategic utilization of AI in industry for further developing the research agenda of VINNOVA’s Strategic Innovation Program Metallic Material. Dr. Mathiason has also been an adjunct member of Jernkontoret’s Board of Digitalization since 2019.
Keywords: “Complex systems, Analytics, Machine Learning”