Oscillatory Neural Networks for Edge AI Computing

Delacour, Corentin; Carapezzi, Stefania; Abernot, Madeleine; Boschetto, Gabriele; Azémard, Nadine; Salles, Jérémie; Gil, Thierry; Todri‐Sanial, Aida

doi:10.1109/isvlsi51109.2021.00066

Cited by 16 publications

(19 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The technology behind Edge AI is a combination of hardware and software that enables AI algorithms to run on edge devices such as smartphones, drones, robots, and other IoT devices [200]. Here are some of the key technologies that are used in Edge AI.…”

Section: Key Technologies For Edge Aimentioning

confidence: 99%

Edge AI: A survey

Singh

Gill

2023

Internet of Things and Cyber-Physical Systems

View full text Add to dashboard Cite

Section: Key Technologies For Edge Aimentioning

confidence: 99%

Edge AI: A survey

Singh

Gill

2023

Internet of Things and Cyber-Physical Systems

View full text Add to dashboard Cite

“…Further investigation and experimentation are needed to report on ONN-HAM analog computation and energy efficiency. However, research around low power analog ONN from novel materials, to novel devices, and up to circuit architectures encourages new system-level architectures and applications exploration [24,22,6,7,8,4].…”

Section: Computation Advantagesmentioning

confidence: 99%

“…Oscillatory Neural Networks (ONNs) [22,6,26,7] are a novel neuromorphic computing paradigm based on coupled oscillators to mimic brain waves observable on electroencephalogram (EEG) [19]. Information is represented in the phase relation between oscillators to limit voltage amplitude and allow low-power computation [24].…”

Section: Introductionmentioning

confidence: 99%

Oscillatory Neural Network as Hetero-Associative Memory for Image Edge Detection

Abernot¹,

Gil²,

Todri‐Sanial³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

The increasing amount of data to be processed on edge devices, such as cameras, has motivated Artificial Intelligence (AI) integration at the edge. Typical image processing methods performed at the edge, such as feature extraction or edge detection, use convolutional filters that are energy, computation, and memory hungry algorithms. But edge devices and cameras have scarce computational resources, bandwidth, and power and are limited due to privacy constraints to send data over to the cloud. Thus, there is a need to process image data at the edge. Over the years, this need has incited a lot of interest in implementing neuromorphic computing at the edge. Neuromorphic systems aim to emulate the biological neural functions to achieve energy-efficient computing.Recently, Oscillatory Neural Networks (ONN) present a novel brain-inspired computing approach by emulating brain oscillations to perform autoassociative memory types of applications. To speed up image edge detection and reduce its power consumption, we perform an in-depth investigation with ONNs. We propose a novel image processing method by using ONNs as a hetero-associative memory (HAM) for image edge detection. We simulate our ONN-HAM solution using first, a Matlab emulator, and then a fully digital ONN design. We show results on gray scale square evaluation maps, also on black and white and gray scale 28x28 MNIST images and finally on black and white 512x512 standard test images. We compare our solution with standard edge detection filters such as Sobel and Canny. Finally, using the fully digital design simulation results, we report on timing and resource characteristics, and evaluate its feasibility for real-time image processing applications. Our digital ONN-HAM solution can process images with up to 120x120 pixels (166 MHz system frequency) respecting real-time camera constraints. This work is the first to explore ONNs as hetero-associative memory for image processing applications.

show abstract

“…In ONNs, information is encoded in the phase difference between oscillators, which also reduces signal voltage amplitude and, consequently, reduces power consumption. Currently, ONNs are being explored from various aspects such as materials [14], devices [15][16][17][18], analog circuit design [19,20] to implementation and applications [21][22][23].…”

Section: Introductionmentioning

confidence: 99%

Oscillatory Neural Networks for Obstacle Avoidance on Mobile Surveillance Robot E4

Abernot

Gil

Kurylin³

et al. 2022

2022 International Joint Conference on Neural Networks (IJCNN)

Self Cite

View full text Add to dashboard Cite

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

show abstract

Oscillatory Neural Networks for Edge AI Computing

Cited by 16 publications

References 25 publications

Edge AI: A survey

Edge AI: A survey

Oscillatory Neural Network as Hetero-Associative Memory for Image Edge Detection

Oscillatory Neural Networks for Obstacle Avoidance on Mobile Surveillance Robot E4

Contact Info

Product

Resources

About