Hendrik Borras, M.Sc.
Hendrik Borras is a doctoral student at the Computing Systems Group at the Institute of Computer Engineering at Heidelberg University. His work is primarily concerned with novel hardware architectures for machine learning outside of the digital domain.
The first time he came in contact with analog technology, was at the CosmicPi Project in 2017. Where he co-designed an analog readout system for the detection of cosmic muons, which as a topic carried over into his bachelor thesis (pdf). Ever since he has had an interest in low-level hardare, be it digital or analog. And he found his more recent curiosity in Machine Learning in 2020, when working on his master thesis in the Computing Systems Group of Prof. Holger Fröning (pdf). After a six month stint in Dublin at the Xilinx Research Labs in 2021 he then started as a Ph.D. candidate back in Heidelberg. And has been working on novel hardware architectures for Bayesian Neural Networks and Deep Neural Networks ever since.
Cheers, Hendrik
P.S.: This summary is somewhat over dramatized, but factually correct.
Research interests
- Novel hardware architectures for Deep and Bayesian Neural Networks
- Noisy hardware, such as photonic and electric analog devices
- High-throughput and low-latency inference of quantized Deep Neural Networks on FPGAs
Curriculum vitae
- Since 2022, PhD candidate in Computer Engineering, Heidelberg University (Germany)
- 2021-2021. Research intern at Xilinx Research Labs Dublin, now AMD RADICAL (Ireland)
- Most of my work done in this context is open-source on Github.
- 2018-2021, Masters in Physics, Heidelberg University (Germany)
- Thesis: Exploring Structured Sparsity within Data-Flow Architecture on Reconfigurable Hardware, pdf
- 2014-2018, Bachelor in Physics, Heidelberg University (Germany)
- Thesis: Determination of the angular distribution of cosmic ray muons and Development of a low-cost silicon detector, pdf
Recent Service (4-year horizon)
Reviewer or Subreviewer
- 2024
- ACM/IEEE International Conference on Computer-Aided Design (ICCAD)
- IEEE Cluster
- IEEE International Conference on Computer Design (ICCD)
- 34th International Conference on Field-Programmable Logic and Applications (FPL)
- International Conference on Supercomputing (ICS)
- 2023
- International Conference on Parallel Processing (ICPP)
- IoT, Edge, and Mobile for Embedded Machine Learning (ITEM) 2023, colocated with ECML-PKDD 2023
Recent Teaching (4-year horizon)
- Winter term 2024
- Organizer and lecturer; undergraduate practical “Neural Networks From Scratch”
- Summer term 2024
- Co-organizer; Seminar: Aktuelle Themen der technischen Informatik
- Summer term 2023
- Teaching Assistant; graduate course “Embedded Machine Learning”
- Winter term 2020/21
- Teaching Assistant; graduate course “GPU computing”
Publications
- Walking Noise: On Layer-Specific Robustness of Neural Architectures against Noisy Computations and Associated Characteristic Learning DynamicsEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD), 2024
@inproceedings{borras2024, title = {Walking Noise: On Layer-Specific Robustness of Neural Architectures against Noisy Computations and Associated Characteristic Learning Dynamics}, author = {Borras, Hendrik and Klein, Bernhard and Fr{\"{o}}ning, Holger}, booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases}, year = {2024}, series = {ECML-PKDD}, url = {https://doi.org/10.1007/978-3-031-70359-1_3}, }
- Probabilistic Photonic Computing with Chaotic LightCoRR, abs/2401.17915, 2024
@article{brckerhoffplckelmann2024probabilistic, title = {Probabilistic Photonic Computing with Chaotic Light}, author = {Brückerhoff-Plückelmann, Frank and Borras, Hendrik and Klein, Bernhard and Varri, Akhil and Becker, Marlon and Dijkstra, Jelle and Brückerhoff, Martin and Wright, C. David and Salinga, Martin and Bhaskaran, Harish and Risse, Benjamin and Fr{\"o}ning, Holger and Pernice, Wolfram}, year = {2024}, volume = {abs/2401.17915}, journal = {CoRR}, url = {https://arxiv.org/abs/2401.17915}, }
- QONNX: Representing Arbitrary-Precision Quantized Neural NetworksCoRR, abs/2206.07527, 2022
@article{DBLP:journals/corr/abs-2206-07527, author = {Pappalardo, Alessandro and Umuroglu, Yaman and Blott, Michaela and Mitrevski, Jovan and Hawks, Benjamin and Tran, Nhan and Loncar, Vladimir and Summers, Sioni and Borras, Hendrik and Muhizi, Jules and Trahms, Matthew and Hsu, Shih{-}Chieh and Hauck, Scott and Duarte, Javier M.}, title = {{QONNX:} Representing Arbitrary-Precision Quantized Neural Networks}, journal = {CoRR}, volume = {abs/2206.07527}, year = {2022}, url = {https://arxiv.org/abs/2206.07527}, doi = {10.48550/ARXIV.2206.07527}, eprinttype = {arXiv}, eprint = {2206.07527}, timestamp = {Tue, 21 Jun 2022 17:35:15 +0200}, }
- Open-source FPGA-ML codesign for the MLPerf Tiny BenchmarkCoRR, abs/2206.11791, 2022
@article{DBLP:journals/corr/abs-2206-11791, author = {Borras, Hendrik and Guglielmo, Giuseppe Di and Duarte, Javier M. and Ghielmetti, Nicol{\`{o}} and Hawks, Benjamin and Hauck, Scott and Hsu, Shih{-}Chieh and Kastner, Ryan and Liang, Jason and Meza, Andres and Muhizi, Jules and Nguyen, Tai and Roy, Rushil and Tran, Nhan and Umuroglu, Yaman and Weng, Olivia and Yokuda, Aidan and Blott, Michaela}, title = {Open-source {FPGA-ML} codesign for the MLPerf Tiny Benchmark}, journal = {CoRR}, volume = {abs/2206.11791}, year = {2022}, url = {https://arxiv.org/abs/2206.11791}, doi = {10.48550/ARXIV.2206.11791}, eprinttype = {arXiv}, eprint = {2206.11791}, timestamp = {Wed, 29 Jun 2022 11:10:54 +0200}, }
Hobbies
- Home automation
- In particular enviromental control of my ultra humid flat.
- Gaming (Valheim, DRG, Ghostrunner, LoL, Nier, Beat Saber, Valorant, …)
- Tinkering with any digital or physical technology that peaks my interest.
- Swimming and diving (if I ever find the time…)