NeuroAdaptiveNet: A Reconfigurable FPGA-Based Neural Network System with Dynamic Model Selection

This paper presents NeuroAdaptiveNet, an FPGA-based neural network framework that dynamically self-adjusts its architectural configurations in real time to maximize performance across diverse datasets. The core innovation is a Dynamic Classifier Selection mechanism, which harnesses the k-Nearest Cen...

Full description

Saved in:
Bibliographic Details
Main Authors: Achraf El Bouazzaoui, Omar Mouhib, Abdelkader Hadjoudja
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Chips
Subjects:
Online Access:https://www.mdpi.com/2674-0729/4/2/24
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849433750644981760
author Achraf El Bouazzaoui
Omar Mouhib
Abdelkader Hadjoudja
author_facet Achraf El Bouazzaoui
Omar Mouhib
Abdelkader Hadjoudja
author_sort Achraf El Bouazzaoui
collection DOAJ
description This paper presents NeuroAdaptiveNet, an FPGA-based neural network framework that dynamically self-adjusts its architectural configurations in real time to maximize performance across diverse datasets. The core innovation is a Dynamic Classifier Selection mechanism, which harnesses the k-Nearest Centroid algorithm to identify the most competent neural network model for each incoming data sample. By adaptively selecting the most suitable model configuration, NeuroAdaptiveNet achieves significantly improved classification accuracy and optimized resource usage compared to conventional, statically configured neural networks. Experimental results on four datasets demonstrate that NeuroAdaptiveNet can reduce FPGA resource utilization by as much as 52.85%, increase classification accuracy by 4.31%, and lower power consumption by up to 24.5%. These gains illustrate the clear advantage of real-time, per-input reconfiguration over static designs. These advantages are particularly crucial for edge computing and embedded applications, where computational constraints and energy efficiency are paramount. The ability of NeuroAdaptiveNet to tailor its neural network parameters and architecture on a per-input basis paves the way for more efficient and accurate AI solutions in resource-constrained environments.
format Article
id doaj-art-18e8bcf7d4864023b4b4c1329d4cb43c
institution Kabale University
issn 2674-0729
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Chips
spelling doaj-art-18e8bcf7d4864023b4b4c1329d4cb43c2025-08-20T03:26:56ZengMDPI AGChips2674-07292025-05-01422410.3390/chips4020024NeuroAdaptiveNet: A Reconfigurable FPGA-Based Neural Network System with Dynamic Model SelectionAchraf El Bouazzaoui0Omar Mouhib1Abdelkader Hadjoudja2SETIME Laboratory, Faculty of Sciences, Ibn Tofail University, Kenitra 14000, MoroccoSETIME Laboratory, Faculty of Sciences, Ibn Tofail University, Kenitra 14000, MoroccoSETIME Laboratory, Faculty of Sciences, Ibn Tofail University, Kenitra 14000, MoroccoThis paper presents NeuroAdaptiveNet, an FPGA-based neural network framework that dynamically self-adjusts its architectural configurations in real time to maximize performance across diverse datasets. The core innovation is a Dynamic Classifier Selection mechanism, which harnesses the k-Nearest Centroid algorithm to identify the most competent neural network model for each incoming data sample. By adaptively selecting the most suitable model configuration, NeuroAdaptiveNet achieves significantly improved classification accuracy and optimized resource usage compared to conventional, statically configured neural networks. Experimental results on four datasets demonstrate that NeuroAdaptiveNet can reduce FPGA resource utilization by as much as 52.85%, increase classification accuracy by 4.31%, and lower power consumption by up to 24.5%. These gains illustrate the clear advantage of real-time, per-input reconfiguration over static designs. These advantages are particularly crucial for edge computing and embedded applications, where computational constraints and energy efficiency are paramount. The ability of NeuroAdaptiveNet to tailor its neural network parameters and architecture on a per-input basis paves the way for more efficient and accurate AI solutions in resource-constrained environments.https://www.mdpi.com/2674-0729/4/2/24FPGAneural networkdynamic classifier selectionneural network acceleratoradaptive accelerator
spellingShingle Achraf El Bouazzaoui
Omar Mouhib
Abdelkader Hadjoudja
NeuroAdaptiveNet: A Reconfigurable FPGA-Based Neural Network System with Dynamic Model Selection
Chips
FPGA
neural network
dynamic classifier selection
neural network accelerator
adaptive accelerator
title NeuroAdaptiveNet: A Reconfigurable FPGA-Based Neural Network System with Dynamic Model Selection
title_full NeuroAdaptiveNet: A Reconfigurable FPGA-Based Neural Network System with Dynamic Model Selection
title_fullStr NeuroAdaptiveNet: A Reconfigurable FPGA-Based Neural Network System with Dynamic Model Selection
title_full_unstemmed NeuroAdaptiveNet: A Reconfigurable FPGA-Based Neural Network System with Dynamic Model Selection
title_short NeuroAdaptiveNet: A Reconfigurable FPGA-Based Neural Network System with Dynamic Model Selection
title_sort neuroadaptivenet a reconfigurable fpga based neural network system with dynamic model selection
topic FPGA
neural network
dynamic classifier selection
neural network accelerator
adaptive accelerator
url https://www.mdpi.com/2674-0729/4/2/24
work_keys_str_mv AT achrafelbouazzaoui neuroadaptivenetareconfigurablefpgabasedneuralnetworksystemwithdynamicmodelselection
AT omarmouhib neuroadaptivenetareconfigurablefpgabasedneuralnetworksystemwithdynamicmodelselection
AT abdelkaderhadjoudja neuroadaptivenetareconfigurablefpgabasedneuralnetworksystemwithdynamicmodelselection