Victor Jean-Baptiste Jung

Victor Jean-Baptiste Jung

Auto-generated profile, based on publicly available information

Victor Jean-Baptiste Jung

Integrated Systems Laboratory, ETH Zrich, Zrich, Switzerland

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
16 May 2024 in IEEE Transactions on Biomedical Circuits and Systems
Verified
Journal Article
Open access
PBPaola Busia
Paola Busia
MSMatteo Antonio Scrugli
Matteo Antonio Scrugli
VJVictor Jean-Baptiste Jung
Victor Jean-Baptiste Jung
LBLuca Benini
Luca Benini
PMPaolo Meloni
Paolo Meloni
Wearable systems for the continuous and real-time monitoring of cardiovascular diseases are becoming widespread and valuable assets in diagnosis and therapy. A promising approach for real-time analysis of the electrocardiographic (ECG) signal and the detection of heart conditions, such as arrhythmia, is represented by the transformer machine learning model. Transformers are powerful models for the classification of time series, although efficient implementation in the wearable domain raises significant design challenges, to combine adequate accuracy and a suitable complexity. In this work, we present a tiny transformer model for the analysis of the ECG signal, requiring only 6k parameters and reaching 98.97% accuracy in the recognition of the 5 most common arrhythmia classes from the MIT-BIH Arrhythmia database, assessed considering 8-bit integer inference as required for efficient execution on low-power microcontroller-based devices. We explored an augmentation-based training approach for improving the robustness against electrode motion artifacts noise, resulting in a worst-case post-deployment performance assessment of 98.36% accuracy. Suitability for wearable monitoring solutions is finally demonstrated through efficient deployment on the parallel ultra-low-power GAP9 processor, where inference execution requires 4.28ms and 0.09mJ.

6 November 2024 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
MSMoritz Scherer
Moritz Scherer
LMLuka Macan
Luka Macan
VJVictor J. B. Jung
Victor J. B. Jung
PWPhilip Wiese
Philip Wiese
LBLuca Bompani
Luca Bompani
LBLuca Benini
Luca Benini
With the rise of embodied foundation models (EFMs), most notably small language models (SLMs), adapting Transformers for the edge applications has become a very active field of research. However, achieving the end-to-end deployment of SLMs on the microcontroller (MCU)-class chips without high-bandwidth off-chip main memory access is still an open challenge. In this article, we demonstrate high efficiency end-to-end SLM deployment on a multicore RISC-V (RV32) MCU augmented with ML instruction extensions and a hardware neural processing unit (NPU). To automate the exploration of the constrained, multidimensional memory versus computation tradeoffs involved in the aggressive SLM deployment on the heterogeneous (multicore+NPU) resources, we introduce Deeploy, a novel deep neural network (DNN) compiler, which generates highly optimized C code requiring minimal runtime support. We demonstrate that Deeploy generates the end-to-end code for executing SLMs, fully exploiting the RV32 cores’ instruction extensions and the NPU. We achieve leading-edge energy and throughput of $490 \; \mu $ J per token, at 340 token per second for an SLM trained on the TinyStories dataset, running for the first time on an MCU-class device without the external memory.

18 November 2024 in IEEE Transactions on Computers
Verified
Journal Article
Closed access
VJVictor Jean-Baptiste Jung
Victor Jean-Baptiste Jung
ABAlessio Burrello
Alessio Burrello
MSMoritz Scherer
Moritz Scherer
FCFrancesco Conti
Francesco Conti
LBLuca Benini
Luca Benini
Transformer networks are rapidly becoming State of the Art (SotA) in many fields, such as Natural Language Processing (NLP) and Computer Vision (CV). Similarly to Convolutional Neural Networks (CNNs), there is a strong push for deploying Transformer models at the extreme edge, ultimately fitting the tiny power budget and memory footprint of Micro-Controller Units (MCUs). However, the early approaches in this direction are mostly ad-hoc, platform, and model-specific. This work aims to enable and optimize the flexible, multi-platform deployment of encoder Tiny Transformers on commercial MCUs. We propose a complete framework to perform end-to-end deployment of Transformer models onto single and multi-core MCUs. Our framework provides an optimized library of kernels to maximize data reuse and avoid unnecessary data marshaling operations into the crucial attention block. A novel Multi-Head Self-Attention (MHSA) inference schedule, named Fused-Weight Self-Attention (FWSA), is introduced, fusing the linear projection weights offline to further reduce the number of operations and parameters. Furthermore, to mitigate the memory peak reached by the computation of the attention map, we present a Depth-First Tiling (DFT) scheme for MHSA tailored for cache-less MCU devices that allows splitting the computation of the attention map into successive steps, never materializing the whole matrix in memory. We evaluate our framework on three different MCU classes exploiting ARM and RISC-V Instruction Set Architecture (ISA), namely the STM32H7 (ARM Cortex M7), the STM32L4 (ARM Cortex M4), and GAP9 (RV32IMC-XpulpV2). We reach an average of 4.79 $\times$ and 2.0 $\times$ lower latency compared to SotA libraries CMSIS-NN (ARM) and PULP-NN (RISC-V), respectively. Moreover, we show that our MHSA depth-first tiling scheme reduces the memory peak by up to 6.19 $\times$, while the fused-weight attention can reduce the runtime by 1.53 $\times$, and number of parameters by 25%. Leveraging the optimizations proposed in this work, we run end-to-end inference of three SotA Tiny Transformers for three applications characterized by different input dimensions and network hyperparameters. We report significant improvements across the networks: for instance, when executing a transformer block for the task of radar-based hand-gesture recognition on GAP9, we achieve a latency of $0.14 \textrm{ms}$ and energy consumption of $4.92 \boldsymbol{\mu}\textrm{J}$, 2.32 $\times$ lower than the SotA PULP-NN library on the same platform.

23 January 2024 in IEEE Transactions on Biomedical Circuits and Systems
Verified
Journal Article
Open access
PBPaola Busia
Paola Busia
ACAndrea Cossettini
Andrea Cossettini
TIThorir Mar Ingolfsson
Thorir Mar Ingolfsson
SBSimone Benatti
Simone Benatti
ABAlessio Burrello
Alessio Burrello
LBLuca Benini
Luca Benini
The long-term, continuous analysis of electroencephalography (EEG) signals on wearable devices to automatically detect seizures in epileptic patients is a high-potential application field for deep neural networks, and specifically for transformers, which are highly suited for end-to-end time series processing without handcrafted feature extraction. In this work, we propose a small-scale transformer detector, the EEGformer, compatible with unobtrusive acquisition setups that use only the temporal channels. EEGformer is the result of a hardware-oriented design exploration, aiming for efficient execution on tiny low-power micro-controller units (MCUs) and low latency and false alarm rate to increase patient and caregiver acceptance. Tests conducted on the CHB-MIT dataset show a 20% reduction of the onset detection latency with respect to the state-of-the-art model for temporal acquisition, with a competitive 73% seizure detection probability and 0.15 false-positive-per-hour (FP/h). Further investigations on a novel and challenging scalp EEG dataset result in the successful detection of 88% of the annotated seizure events, with 0.45 FP/h. We evaluate the deployment of the EEGformer on three commercial low-power computing platforms: the single-core Apollo4 MCU and the GAP8 and GAP9 parallel MCUs. The most efficient implementation (on GAP9) results in as low as 13.7 ms and 0.31 mJ per inference, demonstrating the feasibility of deploying the EEGformer on wearable seizure detection systems with reduced channel count and multi-day battery duration.

8 January 2025 in IEEE Design & Test
Verified
Journal Article
Closed access
PWPhilip Wiese
Philip Wiese
GIGamze İslamoğlu
Gamze İslamoğlu
MSMoritz Scherer
Moritz Scherer
LMLuka Macan
Luka Macan
VJVictor Jean-Baptiste Jung
Victor Jean-Baptiste Jung
LBLuca Benini
Luca Benini
Editor’s notes: This article presents a flexible and co-optimized hardware–software architecture template tailored to the demands of attention-based AI workloads in TinyML systems. The article offers a compelling and practical advancement in energy-efficient attention model deployment for TinyML platforms. —Theocharis Theocharides, University of Cyprus, Cyprus —Marian Verhelst, KU Leuven, Belgium —Vijay Janapa Reddy, Harvard University, USA —Evgeni Gousev, Qualcomm, USA

8 August 2024 in arXiv
Verified
Preprint
Open access
PDF file
MSMoritz Scherer
Moritz Scherer
LMLuka Macan
Luka Macan
VJVictor Jung
Victor Jung
PWPhilip Wiese
Philip Wiese
LBLuca Bompani
Luca Bompani
LBLuca Benini
Luca Benini
With the rise of Embodied Foundation Models (EFMs), most notably Small Language Models (SLMs), adapting Transformers for edge applications has become a very active field of research. However, achieving end-to-end deployment of SLMs on microcontroller (MCU)-class chips without high-bandwidth off-chip main memory access is still an open challenge. In this paper, we demonstrate high-efficiency end-to-end SLM deployment on a multicore RISC-V (RV32) MCU augmented with ML instruction extensions and a hardware neural processing unit (NPU). To automate the exploration of the constrained, multi-dimensional memory vs. computation tradeoffs involved in aggressive SLM deployment on heterogeneous (multicore+NPU) resources, we introduce Deeploy, a novel Deep Neural Network (DNN) compiler, which generates highly-optimized C code requiring minimal runtime support. We demonstrate that Deeploy generates end-to-end code for executing SLMs, fully exploiting the RV32 cores' instruction extensions and the NPU: We achieve leading-edge energy and throughput of \SI{490}{\micro\joule \per Token}, at \SI{340}{Token \per \second} for an SLM trained on the TinyStories dataset, running for the first time on an MCU-class device without external memory.