id: 38644
Title: Modeling Nodes and Cells of Neuron-Equivalentors as Accelerators of Equivalental-Convolutional Self-Learning Neural Structures.
Authors: Krasilenko, V.G., Nikitovych, D.V., Lazarev, A.A.
Keywords: accelerator; neural net; convolutional neural network; neuron-equivalentor; current mirror; vector-matrix procedure; equivalental model
Date of publication: 2025-03-28 19:50:34
Last changes: 2025-03-28 19:50:34
Year of publication: 2024
Summary: In the paper, we consider the urgent need to create highly efficient hardware accelerators for machine learning and deep convolutional neural networks (CNNs), for associative memory models, clustering, and pattern recognition. We show a brief overview of our related works the advantages of the equivalent models (EM) for describing and designing bio-inspired systems. The capacity of neural net on the basis of EM and of its modifications is in several times quantity of neurons. Such neuro-paradigms are very perspective for processing, clustering, recognition, storing large size, strongly correlated, highly noised images and creating of uncontrolled learning machine. And since the basic operational functional nodes of EM are such vector- matrix or matrix-tensor procedures with continuous- logical operations as: normalized vector operations "equivalence", "nonequivalence", and etc., we consider in this paper new conceptual approaches to the design of full-scale arrays of such neuron-equivalentors (NEs) with extended functionality, including different activation functions. Our approach is based on the use of analog and mixed (with special coding) methods for implementing the required operations, building NEs (with number of synapsis from 8 up to 128 and more) and their base cells, nodes based on photosensitive elements and CMOS current mirrors. Simulation results show that the efficiency of NEs relative to the energy intensity is estimated at a value of not less than 10^12 an. op. / sec on W and can be increased. The results confirm the correctness of the concept and the possibility of creating NEs and MIMO structures on their basis.
URI: http://repository.vsau.org/repository/getfile.php/38644.pdf
Publication type: Доповідь конференції
Publication: Modeling, Control and Information Technologies 2024-12-07 | Journal article DOI: 10.31713/MCIT.2024.077 URL: https://itconfdoc.nuwm.edu.ua/index.php/ITConf/article/view/535
In the collections :
Published by: Адміністратор
File : 38644.pdf Size : 1715354 byte Format : Adobe PDF Access : For all
| |
|
|
