The generalization error of overparametrized models: Insights from exact asymptotics Understanding machine learning via exactly solvable statistical physics modelsĭiffusion Methods in Manifold and Fibre Bundle LearningĪviv Regev (Broad Institute, MIT/Harvard)ĭesign for Inference and the Power of Random Experiments in Biology On the Effectiveness of Richardson Extrapolation in Machine Learning Reliable Predictions? Equitable Treatment? Some recent progress in predictive inference Overlap Gap Property: a Provable Barrier to Fast Optimization in Probabilistic Combinatorial Structures Statistical and Computational aspects of Wasserstein Barycenters Self-regularizing Property of Nonparametric Maximum Likelihood Estimator in Mixture ModelsĪlgebraic methods in statistics and data analysisĪmin Coja-Oghlan (Goethe University Frankfurt)Ī Mathematical Perspective of Machine Learning Hierarchical community detection by recursive partitioningīackward Feature Correction: How can Deep Learning perform Deep Learning?Įquilibrium Computation and the Foundations of Deep Learning On estimating the mean of a random vectorĮlizaveta (Liza) Levina (University of Michigan)

Robert Nowak (University of Wisconsin-Madison)īanach Space Representer Theorems for Neural Networks How well can we generalize in high dimension?ĭeep (Convolution) Networks from First Principles Inbar Seroussi (Weizmann Institute of Science) Statistical Frameworks for Mapping 3D Shape Variation onto Genotypic and Phenotypic Variation Lorin Crawford (Microsoft Research & Brown University) Statistical Limits for the Matrix Tensor Product You can subscribe to a calendar here, and to our Youtube channel here. MaD seminars are recorded and streamed live. We took inspiration from the fantastic virtual seminar TCS+ from our friends in Theoretical Computer Science

Of Applied Mathematics, Statistics and Machine Learning. The topics will be in line with the physical MaD seminar hosted at the Center for Data Science, which features leading specialists at the interface For the time being it is a virtual seminar series, and depending on logistics and interest potentially evolve to a multi-location seminar once the world returns to normal (with talks hosted in one of the locations, and streamed). We are excited to announce the MaD+ (math and data plus) seminar, jointly organized between NYU and ETH.

Nyu zoom sign in