generative models

Mapping Properties of Neural Networks and Inverse Problems

Prof. Matti Lassas, Applied Mathematics, University of Helsinki, Finland

Feb 12, 18:00 - 19:00

B1 L4 R4214

Neural Networks generative models

This talk will explore the injectivity and universal approximation properties of neural networks, focusing on one-to-one mappings and the approximation of probability measures using compositions of invertible flow networks and injective layers, with applications in inverse problems, imaging, and generative models.