AxTrain：面向硬件的近似推理神经网络训练
AxTrain: Hardware-Oriented Neural Network Training for Approximate Inference
摘要：
神经网络内在的容错性使得近似计算成为提高神经网络推理能量效率的一种很有前途的技术。传统的近似计算侧重于平衡现有预训练网络的效率-准确性权衡，这可能导致次优解决方案。在本文中，我们提出了AxTrain，一个面向硬件的训练框架，以方便近似计算NN推理。具体来说，AxTrain利用了两种正交方法之间的协同作用-一种是主动搜索具有高容错性的网络参数分布，另一种是通过在训练阶段将近似硬件的噪声分布数值化来被动学习弹性权重。在不同数据集上采用近阈值计算和近似乘法策略的实验结果表明，AxTrain能够获得有弹性的神经网络参数，提高系统能效。
英文摘要：
The intrinsic error tolerance of neural network (NN) makes approximate computing a promising technique to improve the energy efficiency of NN inference. Conventional approximate computing focuses on balancing the efficiency-accuracy trade-off for existing pre-trained networks, which can lead to suboptimal solutions. In this paper, we propose AxTrain, a hardware-oriented training framework to facilitate approximate computing for NN inference. Specifically, AxTrain leverages the synergy between two orthogonal methods---one actively searches for a network parameters distribution with high error tolerance, and the other passively learns resilient weights by numerically incorporating the noise distributions of the approximate hardware in the forward pass during the training phase. Experimental results from various datasets with near-threshold computing and approximation multiplication strategies demonstrate AxTrain's ability to obtain resilient neural network parameters and system energy efficiency improvement.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
