Cover Image

Demonstrating analog inference on the BrainScaleS-2 mobile system

We present the BrainScaleS-2 mobile system as a compact analog inference engine based on the BrainScaleS-2 ASIC and demonstrate its capabilities at classifying a medical electrocardiogram dataset. The analog network core of the ASIC is utilized to perform the multiply-accumulate operations of a conv...

Full description

Saved in:
Bibliographic Details
Main Authors: Stradmann, Yannik (Author) , Billaudelle, Sebastian (Author) , Breitwieser, Oliver (Author) , Ebert, Falk (Author) , Emmel, Arne (Author) , Husmann, Dan (Author) , Ilmberger, Joscha (Author) , Müller, Eric (Author) , Spilger, Philipp (Author) , Weis, Johannes (Author) , Schemmel, Johannes (Author)
Format: Article (Journal)
Language:English
Published: 21 September 2022
In: IEEE open journal of circuits and systems
Year: 2022, Volume: 3, Pages: 252-262
ISSN:2644-1225
DOI:10.1109/OJCAS.2022.3208413
Online Access:Verlag, kostenfrei, Volltext: https://doi.org/10.1109/OJCAS.2022.3208413
Verlag, kostenfrei, Volltext: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9896927
Get full text
Author Notes:Yannik Stradmann, Sebastian Billaudelle, Oliver Breitwieser, Falk Leonard Ebert, Arne Emmel, Dan Husmann, Joscha Ilmberger, Eric Müller, Philipp Spilger, Johannes Weis, and Johannes Schemmel (Member, IEEE)
Description
Summary:We present the BrainScaleS-2 mobile system as a compact analog inference engine based on the BrainScaleS-2 ASIC and demonstrate its capabilities at classifying a medical electrocardiogram dataset. The analog network core of the ASIC is utilized to perform the multiply-accumulate operations of a convolutional deep neural network. At a system power consumption of 5.6W, we measure a total energy consumption of 192 μJ for the ASIC and achieve a classification time of 276 μs per electrocardiographic patient sample. Patients with atrial fibrillation are correctly identified with a detection rate of (93.7 ± 0.7)% at (14.0 ± 1.0)% false positives. The system is directly applicable to edge inference applications due to its small size, power envelope, and flexible I/O capabilities. It has enabled the BrainScaleS-2 ASIC to be operated reliably outside a specialized lab setting. In future applications, the system allows for a combination of conventional machine learning layers with online learning in spiking neural networks on a single neuromorphic platform.
Item Description:"Date of current version 24 October 2022"
Gesehen am 29.09.2022
Physical Description:Online Resource
ISSN:2644-1225
DOI:10.1109/OJCAS.2022.3208413

Similar Items