Energy-Accuracy Scalable Deep Convolutional Neural Networks: A Pareto Analysis - VLSI-SoC: Design and Engineering of Electronics Systems Based on New Computing Paradigms Access content directly
Conference Papers Year : 2019

Energy-Accuracy Scalable Deep Convolutional Neural Networks: A Pareto Analysis

Valentino Peluso
  • Function : Author
  • PersonId : 1025769
Andrea Calimera
  • Function : Author
  • PersonId : 1025753


This work deals with the optimization of Deep Convolutional Neural Networks (ConvNets). It elaborates on the concept of Adaptive Energy-Accuracy Scaling through multi-precision arithmetic, a solution that allows ConvNets to be adapted at run-time and meet different energy budgets and accuracy constraints. The strategy is particularly suited for embedded applications made run at the “edge” on resource-constrained platforms. After the very basics that distinguish the proposed adaptive strategy, the paper recalls the software-to-hardware vertical implementation of precision scalable arithmetic for ConvNets, then it focuses on the energy-driven per-layer precision assignment problem describing a meta-heuristic that searches for the most suited representation of both weights and activations of the neural network. The same heuristic is then used to explore the optimal trade-off providing the Pareto points in the energy-accuracy space. Experiments conducted on three different ConvNets deployed in real-life applications, i.e. Image Classification, Keyword Spotting, and Facial Expression Recognition, show adaptive ConvNets reach better energy-accuracy trade-off w.r.t. conventional static fixed-point quantization methods.
Fichier principal
Vignette du fichier
485996_1_En_6_Chapter.pdf (742.57 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-02321763 , version 1 (21-10-2019)





Valentino Peluso, Andrea Calimera. Energy-Accuracy Scalable Deep Convolutional Neural Networks: A Pareto Analysis. 26th IFIP/IEEE International Conference on Very Large Scale Integration - System on a Chip (VLSI-SoC), Oct 2018, Verona, Italy. pp.107-127, ⟨10.1007/978-3-030-23425-6_6⟩. ⟨hal-02321763⟩
40 View
32 Download



Gmail Facebook X LinkedIn More