Arkadiusz Lewicki, PhD from the Department of Applications of Information Systems took part in one of the largest and most important conferences in the field of artificial intelligence not only in Europe but also in the world – the European Conference on Artificial Intelligence. The event took place from September 30 to October 4 this year. in Cracow.
Arkadiusz Lewicki presented the results of research carried out together with Jagiellonian University scientists on the use of multi-instance learning techniques to solve the problem of interpretability of predictive decisions of machine learning models, taking into account the analysis of unlabeled data. As the research has shown, the problem related to the inability to indicate and assign a specific label to a single instance can be solved by using an aggregation technique using deep neural networks and Bernstein polynomial estimation.
The presented concept related to the scientific article „ProMIL: Probabilistic Multiple Instance Learning for Medical Imaging” was associated with great interest and discussion among experts. The value of this achievement is also proven by the fact that as many as 2,000 applications were sent to the ECAI 23 conference and only 24% of the best and most interesting ones were accepted.
The European Conference on Artificial Intelligence allowed experts to exchange knowledge and experiences, analyze the latest trends and discoveries, access new technologies and tools in the field of artificial intelligence, as well as acquire new contacts, and establish scientific and research cooperation in future projects. Over 1,000 scientists and AI experts from as many as 44 countries participated in this prestigious event sponsored by the European Association for Artificial Intelligence (EurAI), which took place in Poland (Kraków) for the first time in 40 years.
