What question did this study set out to answer?

The central aim is to benchmark the energy efficiency of various convolutional neural network architectures for edge computing in image classification tasks.

April 19, 2026InformationOpen Access

Towards Sustainable AI: Benchmarking Energy Efficiency of Deep Neural Networks for Resource-Constrained Edge Devices

View Full Paper

Ask AI

Bookmark

The central aim is to benchmark the energy efficiency of various convolutional neural network architectures for edge computing in image classification tasks.
Employed a Deep Green Framework for energy-efficient predictions.
Performed benchmarking on CIFAR-10 and CIFAR-100 datasets.
Trained fifteen variants of convolutional neural networks categorized as light-weight and heavy-weight.
Evaluated models based on energy footprint, latency, memory usage, and accuracy metrics.
MobileNetV3-Small achieved the best performance among the models evaluated.
It demonstrated low task latency with high energy efficiency.
The study highlights the suitability of certain models for resource-scarce environments.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Benchmarks energy efficiency in image classification using deep learning models on edge devices, suggesting improvements for resource-limited settings.

Cite This Study

Qamar et al. (2026) studied this question.

synapsesocial.com/papers/69e471c5010ef96374d8e15e https://doi.org/10.3390/info17040380

View Full Paper

Ask AI

Bookmark

Implication

Benchmarks energy efficiency in image classification using deep learning models on edge devices, suggesting improvements for resource-limited settings.

Towards Sustainable AI: Benchmarking Energy Efficiency of Deep Neural Networks for Resource-Constrained Edge Devices

Key Points

Discussion

Implication

Cite This Study

Towards Sustainable AI: Benchmarking Energy Efficiency of Deep Neural Networks for Resource-Constrained Edge Devices

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider