LGMar 7, 2022

Dynamic ConvNets on Tiny Devices via Nested Sparsity

arXiv:2203.03324v18 citationsh-index: 21
Originality Highly original
AI Analysis

This work addresses efficient inference for IoT edge devices, offering a novel compression and training pipeline that is incremental in improving dynamic model deployment.

The paper tackles the problem of deploying dynamic ConvNets on resource-constrained edge devices by introducing Nested Sparse ConvNets, which achieve comparable accuracy, remarkable storage savings, and high performance on an ARM-M7 MCU, outperforming naive variable-latency solutions and being Pareto optimal in accuracy vs. latency compared to state-of-the-art dynamic strategies.

This work introduces a new training and compression pipeline to build Nested Sparse ConvNets, a class of dynamic Convolutional Neural Networks (ConvNets) suited for inference tasks deployed on resource-constrained devices at the edge of the Internet-of-Things. A Nested Sparse ConvNet consists of a single ConvNet architecture containing N sparse sub-networks with nested weights subsets, like a Matryoshka doll, and can trade accuracy for latency at run time, using the model sparsity as a dynamic knob. To attain high accuracy at training time, we propose a gradient masking technique that optimally routes the learning signals across the nested weights subsets. To minimize the storage footprint and efficiently process the obtained models at inference time, we introduce a new sparse matrix compression format with dedicated compute kernels that fruitfully exploit the characteristic of the nested weights subsets. Tested on image classification and object detection tasks on an off-the-shelf ARM-M7 Micro Controller Unit (MCU), Nested Sparse ConvNets outperform variable-latency solutions naively built assembling single sparse models trained as stand-alone instances, achieving (i) comparable accuracy, (ii) remarkable storage savings, and (iii) high performance. Moreover, when compared to state-of-the-art dynamic strategies, like dynamic pruning and layer width scaling, Nested Sparse ConvNets turn out to be Pareto optimal in the accuracy vs. latency space.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes