Memory Optimization for Energy-Efficient Differentially Private Deep Learning
Memory Optimization for Energy-Efficient Differentially Private Deep Learning
复制标题
DOI：
10.1109/tvlsi.2019.2946128
复制
发表时间：
2020-02
影响因子：
2.8
通讯作者：
J. Edstrom;Hritom Das;Yiwen Xu;Na Gong
中科院分区：
文献类型：
--
作者：
J. Edstrom;Hritom Das;Yiwen Xu;Na Gong
随着物联网（IoT）技术的出现和大量数据的可用性，深度学习已经应用于各种人工智能（AI）应用中。然而，使用物联网边缘设备共享个人数据会对个人隐私带来固有的风险。同时，推理过程中所需的能量和内存资源成为资源有限的物联网边缘设备的约束。本文通过考虑差异高效深度学习系统中的隐私性、准确性和能效权衡，带来内存硬件优化，以满足物联网边缘设备的紧张功耗预算。在详细分析这些特性的基础上，建立了一个整数线性规划（ILP）模型，以最小化均方误差（MSE），从而实现最优的输入数据存储器设计。我们在45 nm CMOS工艺的模拟结果表明，所提出的技术可以使近阈值的节能存储器操作不同的隐私要求，在分类精度下降不到1%。
With the advent of Internet of Things (IoT) technologies and availability of a large amount of data, deep learning has been applied in a variety of artificial intelligence (AI) applications. However, sharing personal data using IoT edge devices carries inherent risks to individual privacy. Meanwhile, the energy and memory resources needed during the inference process become a constraint to the resource-limited IoT edge devices. This article brings memory hardware optimization to meet the tight power budget in IoT edge devices by considering the privacy, accuracy, and power efficiency tradeoff in differentially efficient deep learning systems. Based on a detailed analysis on these characteristics, an integer linear programs (ILP) model is developed to minimize mean square error (MSE), thereby enabling optimal input data memory design. Our simulation results in 45-nm CMOS technology show that the proposed technique can enable near-threshold energy-efficient memory operation for different privacy requirements, with less than 1% degradation in classification accuracy.