BottleNet++: An End-to-End Approach for Feature Compression in Device-Edge Co-Inference Systems
BottleNet++: An End-to-End Approach for Feature Compression in Device-Edge Co-Inference Systems
复制标题
DOI：
10.1109/iccworkshops49005.2020.9145068
复制
发表时间：
2019-10
期刊：
影响因子：
--
通讯作者：
Jiawei Shao;Jun Zhang
中科院分区：
文献类型：
--
作者：
Jiawei Shao;Jun Zhang
各种智能移动应用的出现，要求在资源受限的移动设备上部署强大的深度学习模型。设备-边缘协同推理框架通过在移动设备和边缘计算服务器上拆分神经网络来提供一种有前景的解决方案。为了平衡设备上的计算和通信开销，需要仔细选择分割点，而中间特征需要在传输之前进行压缩。现有的研究将模型拆分、特征压缩和通信的设计解耦，这可能会导致移动设备的过度资源消耗。在本文中，我们介绍了一个端到端的体系结构，称为BottleNet++，它由编码器、不可训练的信道层和解码器组成，以实现更高效的特征压缩和传输。编码器和解码器本质上通过轻量级卷积神经网络(CNN)实现联合信源-信道编码，同时显式地考虑了信道噪声的影响。通过利用深度神经网络中间特征的强稀疏性和容错性，BottleNet++获得了比现有方法更高的压缩比。与仅传输中间数据而不进行特征压缩相比，在加性高斯白噪声信道下，BottleNet++可获得高达倍的带宽压缩，在二进制擦除信道下，可获得高达256×bit的压缩比，而分类精度下降不到2%。
The emergence of various intelligent mobile applications demands the deployment of powerful deep learning models at resource-constrained mobile devices. The device-edge co-inference framework provides a promising solution by splitting a neural network at a mobile device and an edge computing server. In order to balance the on-device computation and the communication overhead, the splitting point needs to be carefully picked, while the intermediate feature needs to be compressed before transmission. Existing studies decoupled the design of model splitting, feature compression, and communication, which may lead to excessive resource consumption of the mobile device. In this paper, we introduce an end-to-end architecture, named BottleNet++, that consists of an encoder, a non-trainable channel layer, and a decoder for more efficient feature compression and transmission. The encoder and decoder essentially implement joint source-channel coding via lightweight convolutional neural networks (CNNs), while explicitly considering the effect of channel noise. By exploiting the strong sparsity and the fault-tolerant property of the intermediate feature in deep neural network (DNNs), BottleNet++ achieves a much higher compression ratio than existing methods. Compared with merely transmitting intermediate data without feature compression, BottleNet++ achieves up to 64× bandwidth reduction over the additive white Gaussian noise channel and up to 256× bit compression ratio in the binary erasure channel, with less than 2% reduction in accuracy of classification.