Nils Hülsmeier

Nils Hülsmeier

M. Sc.

Office: NW1, W2190

Tel: +49 421 218 62549

Fax: +49 421 218 98 62549

Mail: Huelsmeier(at)me.uni-bremen.de

Research:    
- Machine learning and deep learning for digital signal processing
- Efficient hardware architectures for machine learning algorithms
- Alternative number formats for machine learning algorithms
 

Teaching:
- Grundlagen Integrierter Schaltungen (Sommersemester)
- Skriptsprachen und Einführung in maschinelles Lernen (Sommersemester)
- Mikroelektronik Labor (Sommersemester)

Publications:

Hybrid SORN Hardware Accelerator for Support Vector Machines
Nils Hülsmeier, Moritz Bärthel, Jochen Rust, Steffen Paul, Proceedings of the Conference for Next Generation Arithmetic (CoNGA) 2023, Singapore, March 2023.

Fused Three-Input SORN Arithmetic
Moritz Bärthel, Chen Yuxing, Nils Hülsmeier, Jochen Rust, Steffen Paul, Proceedings of the Conference for Next Generation Arithmetic (CoNGA) 2023, Singapore, March 2023.

Hybrid SORN Implementation of k-Nearest Neighbor Algorithm on FPGA
Nils Hülsmeier, Moritz Bärthel, Ludwig Karsthof, Jochen Rust, Steffen Paul, 20th IEEE International New Circuits and Systems Conference (NEWCAS) 2022, Quebec, Canada, June 2022.

On the Implementation of Edge Detection Algorithms with SORN Arithmetic
Moritz Bärthel, Nils Hülsmeier, Jochen Rust, Steffen Paul, Proceedings of the Conference for Next Generation Arithmetic (CoNGA) 2022, Singapore, March 2022.

Towards Dynamic Fault Tolerance for Hardware-Implemented Artificial Neural Networks: A Deep Learning Approach
D. Gregorek, N. Hülsmeier and S. Paul, IEEE International Conference on Robotics and Automation (ICRA) 2021 Workshop, Xian, June 2021

SORN-based Cascade Support Vector Machine
Nils Hülsmeier, Moritz Bärthel, Jochen Rust and Steffen Paul, 28th European Signal Processing Conference (EUSIPCO), Amsterdam, Netherlands, 2020

Stochastic Fourier-Transformation Implementation for PVTAR Fault Effects
Kris Niederkleine, Nils Hülsmeier, Theodor Hillebrand, Jochen Rust, Steffen Paul, 7th International Workshop on Analogue and Mixed-Signal Integrated Circuits for Space Applications (AMICSA 2018), Leuven, Belgium, June 2018.

High Throughput Architecture for Inpainting-Based Recovery of Correlated Neural Signals
Sebastian Schmale, Jochen Rust, Nils Hülsmeier, Heiner Lange, Benjamin Knoop, Steffen Paul, 24nd European Signal Processing Conference (EUSIPCO 2016), Budapest, Hungary, September 2016.