Intermittent Inference with Nonuniformly Compressed Multi-Exit Neural Network for Energy Harvesting Powered Devices
Intermittent Inference with Nonuniformly Compressed Multi-Exit Neural Network for Energy Harvesting Powered Devices
复制标题
DOI：
10.1109/dac18072.2020.9218526
复制
发表时间：
2020-04
期刊：
影响因子：
--
通讯作者：
Yawen Wu;Zhepeng Wang;Zhenge Jia;Yiyu Shi;J. Hu
中科院分区：
文献类型：
--
作者：
Yawen Wu;Zhepeng Wang;Zhenge Jia;Yiyu Shi;J. Hu
这项工作旨在通过将轻量级DNNS部署到EH功率的设备上，以实现持久，事件驱动的感应和决策能力（EH）能力设备。但是，收获的能量通常是弱且不可预测的，甚至轻巧的DNN都需要多个功率周期来完成一个推理。为了消除无限期的长期等待，以积累一个推理并优化准确性，我们开发了一种功率痕量吸引和出口引导的网络压缩算法，以压缩和部署多种外观神经网络，以向EH功率驱动的微控制器（MCUS）和MCUS）和根据可用能量执行期间的选择。与最先进的技术相比，实验结果显示出卓越的准确性和潜伏期。
This work aims to enable persistent, event-driven sensing and decision capabilities for energy-harvesting (EH)-powered devices by deploying lightweight DNNs onto EH-powered devices. However, harvested energy is usually weak and unpredictable and even lightweight DNNs take multiple power cycles to finish one inference. To eliminate the indefinite long wait to accumulate energy for one inference and to optimize the accuracy, we developed a power trace-aware and exit-guided network compression algorithm to compress and deploy multi-exit neural networks to EH-powered microcontrollers (MCUs) and select exits during execution according to available energy. The experimental results show superior accuracy and latency compared with state-of-the-art techniques.