Johannes Maly

Johannes Maly is an Assistant Professor at Ludwig Maximilian University of Munich (LMU) since October 2022, holding the Bavarian AI Chair for Mathematical Foundations of Artificial Intelligence. Previously, he held PostDoc positions at Catholic University of Eichstaett-Ingolstadt (2020-2022) and RWTH Aachen University (2019-2020).

His educational background includes:

• PhD in Mathematics from Technical University of Munich (TUM), 2019 (supervised by Prof. Massimo Fornasier)
• M.Sc. in Mathematics from TUM, 2015
• B.Sc. in Mathematics from TUM, 2013

Prof. Maly's research centers on mathematical data science with core focus areas in covariance estimation, neural network approximation properties, implicit bias of gradient descent, and multi-structured signal recovery. A unifying theme across his work is the theoretical investigation of coarse quantization effects, building on his foundational contributions to compressed sensing and extending into modern AI architectures.

Analysis of his recent publications reveals consistent exploration of quantization constraints in high-dimensional statistics and neural network training. Key trends include development of tuning-free covariance estimators using dithering techniques, characterization of implicit regularization in overparameterized models, and novel quantization approaches for neural networks that balance hardware efficiency with theoretical guarantees.

His scientific recognition includes:

• relAI Fellow (Zuse School for Reliable AI)
• MCML Associate (Munich Center for Machine Learning)

Prof. Maly's research is supported through his Bavarian AI Chair appointment and fellowships. He teaches advanced courses including Convex Optimization, High-dimensional Probability, and Mathematical Introduction to Data Science at LMU Munich, with prior teaching roles at Catholic University of Eichstaett-Ingolstadt and RWTH Aachen University. While specific student names are not listed in the source material, his position involves supervision of graduate researchers.

He leads the Mathematical Data Science and Artificial Intelligence research group at LMU Munich under the Bavarian AI Chair, collaborating with MCML and relAI networks. The group investigates theoretical challenges in quantization, low-rank matrix recovery, and optimization dynamics for next-generation AI systems.