Demonstrating Analog Inference on the BrainScaleS-2 Mobile System

03/29/2021
by   Yannik Stradmann, et al.
0

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. We measure a total energy consumption of 192uJ for the ASIC and achieve a classification time of 276us per electrocardiographic patient sample. Patients with atrial fibrillation are correctly identified with a detection rate of 93.7(7) positives. The system is directly applicable to edge inference applications due to its small size, power envelope and flexible I/O capabilities. Possible future applications can furthermore combine conventional machine learning layers with online-learning in spiking neural networks on a single BrainScaleS-2 ASIC. The system has successfully participated and proven to operate reliably in the independently judged competition "Pilotinnovationswettbewerb 'Energieeffizientes KI-System'" of the German Federal Ministry of Education and Research (BMBF).

READ FULL TEXT

page 1

page 2

page 5

research
12/24/2022

Mantis: Enabling Energy-Efficient Autonomous Mobile Agents with Spiking Neural Networks

Autonomous mobile agents such as unmanned aerial vehicles (UAVs) and mob...
research
03/03/2023

TopSpark: A Timestep Optimization Methodology for Energy-Efficient Spiking Neural Networks on Autonomous Mobile Agents

Autonomous mobile agents require low-power/energy-efficient machine lear...
research
10/24/2022

An Analytical Estimation of Spiking Neural Networks Energy Efficiency

Spiking Neural Networks are a type of neural networks where neurons comm...
research
02/21/2022

Energy-Efficient Respiratory Anomaly Detection in Premature Newborn Infants

Precise monitoring of respiratory rate in premature infants is essential...
research
08/25/2022

CMOS-based area-and-power-efficient neuron and synapse circuits for time-domain analog spiking neural networks

Conventional neural structures tend to communicate through analog quanti...
research
09/05/2023

HW/SW Codesign for Robust and Efficient Binarized SNNs by Capacitor Minimization

Using accelerators based on analog computing is an efficient way to proc...
research
12/10/2021

Synchronous Unsupervised STDP Learning with Stochastic STT-MRAM Switching

The use of analog resistance states for storing weights in neuromorphic ...

Please sign up or login with your details

Forgot password? Click here to reset