An Energy-Efficient Inference Method in Convolutional Neural Networks Based on Dynamic Adjustment of the Pruning Level
An Energy-Efficient Inference Method in Convolutional Neural Networks Based on Dynamic Adjustment of the Pruning Level
复制标题
一种基于动态调整剪枝水平的卷积神经网络节能推理方法
DOI：
10.1145/3460972
复制
发表时间：
2021
影响因子：
1.4
通讯作者：
Pedram, Massoud
中科院分区：
文献类型：
--
作者：
Maleki, Mohammad-Ali;Nabipour-Meybodi, Alireza;Kamal, Mehdi;Afzali-Kusha, Ali;Pedram, Massoud
在这篇文章中，我们提出了一种用于图像分类应用中卷积神经网络的低能耗推理方法。如果得到的网络可以提供正确的输出，则通过使用高度修剪（低能量）的网络来实现较低的能量消耗。更具体地说，所提出的推理方法利用两个修剪神经网络（NN），即温和和积极修剪网络，这两个都是离线设计的。在该系统中，第三个NN利用输入数据在线选择适当的修剪网络。第三个网络，其特征提取，采用相同的卷积层的积极修剪NN，从而减少了在线管理的开销。有一些精度损失所提出的方法，其中，对于一个给定的精度水平，所提出的方法的能量增益是相当大的情况下，采用任何一个修剪水平。该方法与修剪方法和网络结构无关。Eyeriss硬件加速器平台上的一些国家的最先进的神经网络架构的建议的推理方法的有效性进行评估。我们的研究表明，这种方法可以提供，平均而言，70%的能量减少相比，在CIFAR-10数据集的成本约3%的准确性损失的原始神经网络。
In this article, we present a low-energy inference method for convolutional neural networks in image classification applications. The lower energy consumption is achieved by using a highly pruned (lower-energy) network if the resulting network can provide a correct output. More specifically, the proposed inference method makes use of two pruned neural networks (NNs), namely mildly and aggressively pruned networks, which are both designed offline. In the system, a third NN makes use of the input data for the online selection of the appropriate pruned network. The third network, for its feature extraction, employs the same convolutional layers as those of the aggressively pruned NN, thereby reducing the overhead of the online management. There is some accuracy loss induced by the proposed method where, for a given level of accuracy, the energy gain of the proposed method is considerably larger than the case of employing any one pruning level. The proposed method is independent of both the pruning method and the network architecture. The efficacy of the proposed inference method is assessed on Eyeriss hardware accelerator platform for some of the state-of-the-art NN architectures. Our studies show that this method may provide, on average, 70% energy reduction compared to the original NN at the cost of about 3% accuracy loss on the CIFAR-10 dataset.
彼得·胡滕洛赫 (1931–2013)
影响因子：
64.8
作者：
C. Walsh
通讯作者：
C. Walsh