To understand AI – Silesian University of Technology is a leader of the Horizon Europe project
How to better understand decisions made by artificial intelligence algorithms in fields such as medicine, industry, or natural language processing, also in the context of cybersecurity – these are the assumptions of the project: “Towards an Understanding of Artificial Intelligence.” Silesian University of Technology hosted doctoral students and scientists from all over the world, involved in design research.
The main challenge and goal of the ‘Towards an Understanding of Artificial Intelligence” TUAI project is to better understand the operation of artificial intelligence systems, which - the more advanced they are, the more often their operation and, for example, the decisions made by algorithms are less understandable to users.
- The point is that AI systems should support people and not suggest decisions that are unfavourable to them. Therefore, the main goal of the project is to better understand the decisions made by algorithms. We do not work on the effectiveness of models - what if their operation remains illegible or prone to errors resulting from improper learning, leading to decisions that are contrary to ethics, common sense or user interest - explains Dr Hab. Eng. Rafał Cupek, Prof SUT from the Faculty of Automatic Control, Electronics and Computer Science, TUAI project coordinator.
The project covers four key areas of application: Medicine, Industry 5.0, Natural Language Processing and Cybersecurity. Each of them presents different challenges – from the need for understandable recommendations for doctors and patients, through the assessment of the impact of AI on industry workers, to counteracting bias in the case of natural language processing (NLP) and protecting systems from cyber attacks.
– The most important thing in this project are people, we focus not only on scientific research, but above all on human beings. The aim of the initiative is to implement a multidisciplinary training program in the field of explainable artificial intelligence for 13 PhD students who gain experience in various research environments – added Professor Cupek.
The key idea of the project is mobility – all participants come from outside the country where they come to study at the host university for three years and participate in a project delegated to partner universities for several months. Silesian University of Technology hosts three PhD students, but this time almost all project participants met at our University. They visited AIUT laboratories and company headquarters and took part in workshops where they worked with AMR autonomous mobile robots. One of the objectives of the project is to develop collision avoidance strategies, create algorithms and test solutions that can improve the way AMRs detect and bypass obstacles in an industrial environment.
– This is a very valuable visit for me, I am particularly impressed by the close cooperation of the Silesian University of Technology with industry, the fact that companies actively support the activities of the university, invest in laboratories. Thanks to this approach, the quality of education increases significantly. I will also add that the infrastructure here is very modern and impressive, - said Juan Camilo España, a PhD student from the University of Oviedo in Spain.
- Our main goal is to develop new deep learning models for time series analysis, including time series grouping and detection of anomalies in the industrial environment. We want to use our knowledge and experience to develop existing solutions. - We work in a variety of areas, not just industrial, but with machine learning techniques, we can create new models and systems that will allow industry to take a step forward and support future generations, - said David Camacho of the Polytechnic University of Madrid.
All participants of the meeting at the Silesian University of Technology emphasized the benefits of mobility, exchange of experiences and thoughts with project participants from different parts of the world.
– Currently, the possibilities of cooperation and foreign trips are very wide. I think it is a great value for us to gain new experience and take advantage of such opportunities, especially in cooperation with industry as in this case. Thanks to this, we do not only learn theory at the university, but also how to implement it in practice and contribute to the improvement of the quality of life and the development of society, - said Dina Tri Utari from Indonesia, who is a PhD student at the Norwegian University of Science and Technology.
The “Towards an Understanding of Artificial Intelligence” TUAI project is implemented by the Silesian University of Technology as part of Maria Skłodowska-Curie actions (MSCA) and is part of the most prestigious science support program in the European Union - Horizon Europe. Our University cooperates with the University of Naples, the Polytechnic University of Madrid, the Norwegian University of Science and Technology, the Norwegian University of Applied Sciences, the University of Oviedo and representatives of industry, the example of which is the AIUT company in Poland. The project will be implemented until the end of September 2028.
