N
Nathan Laubeuf
Researcher at École Polytechnique Fédérale de Lausanne
Publications - 4
Citations - 38
Nathan Laubeuf is an academic researcher from École Polytechnique Fédérale de Lausanne. The author has contributed to research in topics: Computer science & Stretchable electronics. The author has an hindex of 1, co-authored 1 publications receiving 22 citations.
Papers
More filters
Journal ArticleDOI
Gallium‐Based Thin Films for Wearable Human Motion Sensors
TL;DR: In this paper, the surface coating and topography of silicone substrates are engineered to enable precisely defined, micrometer-thick liquid metal patterns over large (>10 cm2) surface areas, and high design versatility.
Journal ArticleDOI
Dynamic Quantization Range Control for Analog-in-Memory Neural Networks Acceleration
Nathan Laubeuf,J. Doevenspeck,Ioannis A. Papistas,Michele Caselli,Stefan Cosemans,Peter Vrancx,D. Bhattacharjee,Arindam Mallik,Peter Debacker,Diederik Verkest,Francky Catthoor,Rudy Lauwereins +11 more
TL;DR: This work demonstrates that dynamic control over this quantizationrange is possible but also desirable for analog neural networks acceleration, and an AiMC compatible quantization flow coupled with a hardware aware quantization range driving technique is introduced to fully exploit these dynamic ranges.
Proceedings ArticleDOI
Learn to Learn on Chip: Hardware-aware Meta-learning for Quantized Few-shot Learning at the Edge
Nitish Satya Murthy,Peter Vrancx,Nathan Laubeuf,Peter Debacker,Francky Catthoor,Marian Verhelst +5 more
TL;DR: In this paper , a modified meta-learning algorithm is proposed to enable quantized fine-tuning to optimally condition the models for on-chip few-shot learning, which involves the inclusion of target hardware constraints upfront in the meta learning process.
Proceedings ArticleDOI
Analog Compute in Memory and Breaking Digital Number Representations
TL;DR: In this paper , the authors show that these lacking performance can be partly attributed to mismatched arithmetic assumption, and they show that correcting those assumptions results in more robust and accurate DNNs on analog platforms.