Sign In

Matthijs Van keirsbilck

Research Scientist at NVIDIA

Matthijs Van Keirsbilck is a Research Scientist at NVIDIA, specializing in the foundations of neural networks and machine learning. His work focuses on efficient neural network architecture design, including leveraging structural sparsity and quantization to optimize performance.14

He holds a Master's degree in Electrical Engineering with a focus on Integrated Circuit Design, having studied at KU Leuven in Belgium and the Technical University of Munich in Germany. His thesis explored robust speech recognition using simultaneous audio and image processing.14

Matthijs has contributed to several notable research projects, including co-authoring the paper "Hymba: A Hybrid-head Architecture for Small Language Models," which proposed innovative techniques for improving the efficiency and performance of small language models.2 Additionally, he has worked on projects like "Artificial Neural Networks Generated by Low Discrepancy Sequences," which explored sparse neural networks for computational efficiency.3

He has also published extensively in the fields of machine learning, deep learning, and efficient computing, with his work cited multiple times.5

Highlights

Nov 20 · arxiv.org
Hymba: A Hybrid-head Architecture for Small Language Models - arXiv
People - Research at NVIDIA

Related Questions

What are some of the key projects Matthijs Van Keirsbilck has worked on at NVIDIA?
How does Matthijs Van Keirsbilck's work on neural network architecture contribute to AI advancements?
What inspired Matthijs Van Keirsbilck to pursue a career in electrical engineering and machine learning?
Can you provide more details about the Hymba architecture and its benefits?
What are the main challenges Matthijs Van Keirsbilck faces in his research on neural networks?
Matthijs Van keirsbilck
Matthijs Van keirsbilck, photo 1
Matthijs Van keirsbilck, photo 2
Add to my network

Location

Berlin, Berlin, Germany