Institution profile

University of Münster

Academic institutioneurope · de
Official website
Research library146linked papers
Opportunities0open roles
Selected work

Representative Papers

Physical Analog Kolmogorov-Arnold Networks based on Reconfigurable Nonlinear-Processing Units

Feb 07, 2026

This work addresses the challenge of efficiently implementing the learnable nonlinear edge functions in Kolmogorov–Arnold Networks (KANs) in hardware by proposing an analog KAN architecture based on a Reconfigurable Nonlinear Processing Unit (RNPU). Leveraging multi-terminal nanosilicon devices that natively support programmable nonlinear transformations, the design employs the RNPU as its fundamental computational element, integrated with analog mixed-signal interfaces to achieve high parameter efficiency, low power consumption, and minimal silicon area for edge neural network deployment. Experimental results demonstrate that, at comparable approximation error levels, the proposed architecture reduces energy consumption by two to three orders of magnitude and chip area by approximately one order of magnitude relative to digital fixed-point MLP implementations, achieving a single-inference energy cost of merely 250 pJ with a latency of about 600 ns.

1 citationsRead paper

Transformer-Based Parameter Fitting of Models Derived from Bloch-McConnell Equations for CEST MRI Analysis

Feb 06, 2026MLMI@MICCAI

This work proposes a novel paradigm for quantitative CEST MRI by introducing, for the first time, a self-supervised Transformer architecture to directly estimate physical parameters—such as metabolite concentration, proton exchange rate, and relaxation rates—from in vitro CEST spectra. These parameters are grounded in the Bloch-McConnell equations, which describe the complex coupling of multiple physiological factors that traditionally hinder accurate CEST signal quantification. Conventional gradient-based optimization methods suffer from low efficiency and susceptibility to local minima, whereas the proposed end-to-end neural network overcomes these limitations. The method demonstrates substantial improvements over classical solvers in both accuracy and computational efficiency, establishing a new framework for robust and rapid CEST parameter mapping.

1 citationsRead paper
Recent publications

Latest Papers