CoopNet: Cooperative Convolutional Neural Network for Low-Power MCUs
This work addresses efficient neural network deployment for resource-constrained edge devices, representing an incremental improvement over existing optimization techniques.
The paper tackled the problem of deploying convolutional neural networks on low-power microcontrollers by proposing CoopNet, a cooperative approach combining fixed-point quantization and binarization, which resulted in lower latency and higher accuracy compared to using these methods separately.
Fixed-point quantization and binarization are two reduction methods adopted to deploy Convolutional Neural Networks (CNN) on end-nodes powered by low-power micro-controller units (MCUs). While most of the existing works use them as stand-alone optimizations, this work aims at demonstrating there is margin for a joint cooperation that leads to inferential engines with lower latency and higher accuracy. Called CoopNet, the proposed heterogeneous model is conceived, implemented and tested on off-the-shelf MCUs with small on-chip memory and few computational resources. Experimental results conducted on three different CNNs using as test-bench the low-power RISC core of the Cortex-M family by ARM validate the CoopNet proposal by showing substantial improvements w.r.t. designs where quantization and binarization are applied separately.