International Journal of Advance Computational Engineering and Networking (IJACEN)
.
Follow Us On :
current issues
Volume-12,Issue-9  ( Sep, 2024 )
Past issues
  1. Volume-12,Issue-8  ( Aug, 2024 )
  2. Volume-12,Issue-7  ( Jul, 2024 )
  3. Volume-12,Issue-6  ( Jun, 2024 )
  4. Volume-12,Issue-5  ( May, 2024 )
  5. Volume-12,Issue-4  ( Apr, 2024 )
  6. Volume-12,Issue-3  ( Mar, 2024 )
  7. Volume-12,Issue-2  ( Feb, 2024 )
  8. Volume-12,Issue-1  ( Jan, 2024 )
  9. Volume-11,Issue-12  ( Dec, 2023 )
  10. Volume-11,Issue-11  ( Nov, 2023 )

Statistics report
Feb. 2025
Submitted Papers : 80
Accepted Papers : 10
Rejected Papers : 70
Acc. Perc : 12%
Issue Published : 141
Paper Published : 1672
No. of Authors : 4423
  Journal Paper


Paper Title :
Enabling On-Line Precision Scaling for Energy-Driven Adaptive ConvNets

Author :Valentino Peluso, Andrea Calimera, Enrico Macii

Article Citation :Valentino Peluso ,Andrea Calimera ,Enrico Macii , (2024 ) " Enabling On-Line Precision Scaling for Energy-Driven Adaptive ConvNets " , International Journal of Advance Computational Engineering and Networking (IJACEN) , pp. 45-51, Volume-12,Issue-9

Abstract : Adaptive Convolutional Deep Neural Networks (Adaptive ConvNets) can reshape their behavior to reach a better trade-off between computational effort and prediction accuracy, which is a key feature for energy-efficient edge applications. As a subclass, energy-driven adaptive ConvNets can self-tune their energy footprint upon request based on an external trigger produced at the application level. This work introduces a design and optimization strategy based on the concept of online precision scaling. The optimization was built and formulated as a multi-objective problem solved via a modified version of the Non-Dominated Sorting Genetic Algorithm II (NSGA-II) that guarantees a fast design space exploration. The simulation results we collected using a software-programmable neural accelerator architecture with mixed-precision arithmetic demonstrate our approach enables ConvNets to shift over more Pareto optimal operating points, with energy savings up to 35% for less than 3% accuracy loss. Keywords - Deep Learning, Optimization, Energy Efficiency

Type : Research paper

Published : Volume-12,Issue-9


DOIONLINE NO - IJACEN-IRAJ-DOIONLINE-21247   View Here

Copyright: © Institute of Research and Journals

| PDF |
Viewed - 10
| Published on 2025-01-08
   
   
IRAJ Other Journals
IJACEN updates
Paper Submission is open now for upcoming Issue.
The Conference World

JOURNAL SUPPORTED BY