Modern Regularization Methods in Inverse Problems and Data Science

Explore recent developments in variational methods for inverse problems and data science in this 45-minute talk by Martin Burger. Delve into the basic properties required for convergent regularization schemes and examine the derivation of quantitative estimates, including the use of Bregman distances for convex variational and iterative methods. Discover the connections between classical regularization theory and machine learning, as the speaker reinterprets machine learning problems within the framework of regularization theory and vice versa. Gain insights into the application of variational methods for inverse problems through the lens of risk minimization. Access the accompanying slides for a visual aid to enhance your understanding of these modern regularization techniques in inverse problems and data science.

Martin Burger: Modern regularization methods in inverse problems and data science