Biophotonics Artificial Intelligence School

In this talk I will introduce modern deep generative models starting from basic concepts such as maximum likelihood learning moving forward to recent models and algorithms to learn to generate data such as Generative Adversarial Networks and Diffusion Models. We will dive into these two models identifying the main issues and advantages. We will finally look at how these models can be applied to solve problems such as image compression, anomaly detection and uncertainty estimation.