Exposé de Masashi Sugiyama

mercredi 17 juillet 2024
Masashi Sugiyama, Professor at the University of Tokyo and Director of the RIKEN Center for Advanced Intelligence Project. He was invited by the MAASAI team and agreed to give a lecture will take place on Wednesday 17 July 2024 at 2pm at Inria, Euler Violet building. The title of his talk is « Machine Learning from Imperfect Information ».
He has made numerous contributions to machine learning algorithms, particularly with regard to problems associated with difficult data (changes in data sets, semi-supervised learning, learning without positive labels, label noise).
His work can be viewed at : https://scholar.google.fr/citations?user=GkYIrlIAAAAJ&hl=en
Abstract :
In recent machine learning applications, training data is often full of uncertainties. In this talk, I will give an overview of our research on reliable machine learning from imperfect information, including weakly supervised learning, noise-robust learning, and transfer learning. Then, I will discuss our recent challenges to integrate these approaches and develop a generic machine learning methodology with fewer modelling assumptions.