Highlights

NumGM project approved

Byadmin 4 September 20267 September 2026

The German Research Foundation has approved the 36-month project Numerically Efficient Learning of Generative Models and Beyond (NumGM), a collaboration between Benjamin Gess and Gabriele Steidl. The project develops geometry-aware generative models and accelerated numerical algorithms designed to make training and inference more stable, efficient and sustainable.

Programme: DFG Research Grants Programme
DFG reference: STE 571/23-1
Project number: 580928721
AOBJ: 722293
Approved duration: 36 months

Project details

Publications

Functional Neural Wavefunction Optimization
Byadmin 14 July 202527 September 2026

We propose a framework for the design and analysis of optimization algorithms in variational quantum Monte Carlo, drawing on geometric insights into the corresponding function space. The framework translates infinite-dimensional optimization dynamics into tractable parameter-space algorithms through a Galerkin projection onto the…

Read More Functional Neural Wavefunction Optimization
Publications

Low temperature expansion for the Euclidean $Φ^4_2$-measure
Byadmin 22 April 20249 October 2026

Published · Transactions of the American Mathematical Society, (2026). We study asymptotic expansions of the Euclidean $Φ^4_2$-measure in the low-temperature regime. In particular, this extends the asymptotic expansions of Gaussian function space integrals developed in Schilder (1966) and Ellis and Rosen…

Read More Low temperature expansion for the Euclidean $Φ^4_2$-measure
Member Publication Archive

Density bounds for solutions to differential equations driven by Gaussian rough paths
Byadmin 7 December 201728 September 2026

Published · Journal of theoretical probability, 33 (2020) 2, pp. 611-648. We consider finite dimensional rough differential equations driven by centered Gaussian processes. Combining Malliavin calculus, rough paths techniques and interpolation inequalities, we establish upper bounds on the density of the corresponding solution…

Read More Density bounds for solutions to differential equations driven by Gaussian rough paths
Publications

THINNs: Thermodynamically Informed Neural Networks
Byadmin 23 September 202527 September 2026

Physics-Informed Neural Networks (PINNs) are a class of deep learning models aiming to approximate solutions of PDEs by training neural networks to minimize the residual of the equation. Focusing on non-equilibrium fluctuating systems, we propose a physically informed choice of penalization that…

Read More THINNs: Thermodynamically Informed Neural Networks
Publications

Large Spikes in Stochastic Gradient Descent: A Large-Deviations View
Byadmin 10 March 202627 September 2026

Large loss spikes in stochastic gradient descent are studied through a rigorous large-deviations analysis for a shallow, fully connected network in the NTK scaling. In contrast to full-batch gradient descent, the catapult phase is shown to split into inflationary and deflationary regimes,…

Read More Large Spikes in Stochastic Gradient Descent: A Large-Deviations View
Member Publication Archive

Convergence rates for the stochastic gradient descent method for non-convex objective functions
Byadmin 2 April 201928 September 2026

Published · Journal of machine learning research, 21 (2020) 136, pp. 1-48. We prove the local convergence to minima and estimates on the rate of convergence for the stochastic gradient descent method in the case of not necessarily globally convex nor contracting objective functions…

Read More Convergence rates for the stochastic gradient descent method for non-convex objective functions

Similar Posts