Steffen Limmer
Steffen Limmer

Research Scientist

Siemens AG

About Me

I am a Research Scientist in the Foundational Technologies Division of Siemens AG. I work on research projects in the area of Physical AI and on developing AI-enhanced products. In 2019, I received a PhD in EECS from TU Berlin (with highest distinction) and my thesis was on Neural Architectures for Sparse Signals.

Interests
  • Signal Processing
  • Machine Learning
  • Reinforcement Learning
  • Physical AI
Education

  • PhD in EECS

    Technical Universitiy of Berlin

  • Dipl.-Ing. in EECS

    Technical University of Munich

Publications
Neural-ANOVA: Analytical Model Decomposition using Automatic Integration
Physics-informed neural networks for pathloss prediction
Physical AI

Physics-informed neural networks for pathloss prediction

This paper presents a novel physics-informed machine learning method for pathloss prediction that significantly enhances generalization and prediction accuracy. The approach uniquely integrates both the inherent physical relationships within the spatial loss field and empirical pathloss measurements directly into the neural network's training process. This dual-constraint learning problem enables the model to achieve superior performance with fewer layers and parameters, resulting in exceptionally fast inference times crucial for subsequent applications like localization. Furthermore, the physics-informed framework substantially reduces the need for extensive training data, making this method highly adaptable and practical for diverse real-world pathloss prediction challenges.

A neural architecture for Bayesian compressive sensing over the simplex via Laplace techniques
Sparse Signal Processing

A neural architecture for Bayesian compressive sensing over the simplex via Laplace techniques

This paper introduces a novel theoretical and conceptual framework for designing neural architectures specifically for Bayesian compressive sensing of simplex-constrained sparse stochastic vectors. The core idea involves reframing the MMSE estimation problem as computing the centroid of a polytope, which is the intersection of a simplex and an affine subspace defined by compressive measurements. Leveraging multidimensional Laplace techniques, the authors derive a closed-form solution for this centroid computation and demonstrate how to directly map this solution to a neural network architecture composed of threshold, ReLU, and rectified polynomial activation functions. This unique construction results in an architecture where the number of layers equals the number of measurements, offering faster solutions in low-measurement scenarios and exhibiting robustness to small model mismatches. Simulations further indicate that this proposed architecture achieves superior approximations with fewer parameters compared to standard ReLU networks in supervised learning contexts.

Towards optimal nonlinearities for sparse recovery using higher-order statistics
Sparse Signal Processing

Towards optimal nonlinearities for sparse recovery using higher-order statistics

This paper investigates machine learning techniques to achieve low-latency approximate solutions for inverse problems, specifically focusing on recovering sparse stochastic signals within lp​-balls using a probabilistic framework. The authors analyze the Bayesian mean-square-error (MSE) for two estimators. a linear one, and a structured nonlinear one comprising a linear operator followed by a Cartesian product of univariate nonlinear mappings. Crucially, the proposed nonlinear estimator maintains comparable complexity to its linear counterpart due to the efficient hardware implementation of the nonlinear mapping via look-up tables (LUTs). This structure is well-suited for neural networks and single-iterate shrinkage/thresholding algorithms, and an alternating minimization technique yields optimized operators and mappings that converge in MSE, making it highly appealing for real-time applications where traditional iterative optimization is infeasible.

A simple algorithm for approximation by nomographic functions
Talks