Enabling binary neural network training on the edge
File(s)
Published version
Author(s)
Type
Journal Article
Abstract
The ever-growing computational demands of increasingly complex machine learning models frequently necessitate the use of powerful cloud-based infrastructure for their training. Binary neural networks are known to be promising candidates for on-device inference due to their extreme compute and memory savings over higher-precision alternatives. However, their existing training methods require the concurrent storage of high-precision activations for all layers, generally making learning on memory-constrained devices infeasible. In this article, we demonstrate that the backward propagation operations needed for binary neural network training are strongly robust to quantization, thereby making on-the-edge learning with modern models a practical proposition. We introduce a low-cost binary neural network training strategy exhibiting sizable memory footprint reductions while inducing little to no accuracy loss vs Courbariaux & Bengio’s standard approach. These decreases are primarily enabled through the retention of activations exclusively in binary format. Against the latter algorithm, our drop-in replacement sees memory requirement reductions of 3–5×, while reaching similar test accuracy (± 2 pp) in comparable time, across a range of small-scale models trained to classify popular datasets. We also demonstrate from-scratch ImageNet training of binarized ResNet-18, achieving a 3.78× memory reduction. Our work is open-source, and includes the Raspberry Pi-targeted prototype we used to verify our modeled memory decreases and capture the associated energy drops. Such savings will allow for unnecessary cloud offloading to be avoided, reducing latency, increasing energy efficiency, and safeguarding end-user privacy.
Date Issued
2023-11
Date Acceptance
2023-09-11
Citation
ACM Transactions on Embedded Computing Systems, 2023, 22 (6), pp.1-19
ISSN
1539-9087
Publisher
Association for Computing Machinery (ACM)
Start Page
1
End Page
19
Journal / Book Title
ACM Transactions on Embedded Computing Systems
Volume
22
Issue
6
Copyright Statement
Copyright This paper is embargoed until publication. Once published the Author Accepted Manuscript (AAM) will be available on immediate open access with a CC-BY License.
License URL
Identifier
https://dl.acm.org/doi/10.1145/3626100
Subjects
CCS Concepts: • Computing methodologies → Machine learning
• Computer systems organization → Embedded systems
Deep neural network, binary neural network, training, edge devices, embedded systems, memory reduction
Publication Status
Published
Coverage Spatial
Honolulu, Hawaii
Article Number
105
Date Publish Online
2023-11-09