At AIST, machine learning and data science means preparing and analysing real-world data so that it leads to sound decisions. Projects range from AI-supported knowledge bases for new production lines in the plastics industry and predictive maintenance for vehicles (AAOS Eco+Maintenance) to reducing alarm floods in critical infrastructure (iReduce) and automatically checking the coherence of survey results.
The group uses large language models where they relieve experts of routine work – generating learning material (Soggiorno), answering social media comments (Replient) or drafting presentations. At the same time, AIST works on making AI explainable: in the FWF-funded doctoral programme Human-Centered Artificial Intelligence together with Johannes Kepler University Linz, and in the MINERVA project, which combines neural perception with symbolic reasoning.
