当前位置: X-MOL 学术IEEE Trans. Biomed. Circuits Syst. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Application of Deep Compression Technique in Spiking Neural Network Chip.
IEEE Transactions on Biomedical Circuits and Systems ( IF 5.1 ) Pub Date : 2019-11-11 , DOI: 10.1109/tbcas.2019.2952714
Yanchen Liu , Kun Qian , Shaogang Hu , Kun An , Sheng Xu , Xitong Zhan , J. J. Wang , Rui Guo , Yuancong Wu , Tu-Pei Chen , Qi Yu , Yang Liu

In this paper, a reconfigurable and scalable spiking neural network processor, containing 192 neurons and 6144 synapses, is developed. By using deep compression technique in spiking neural network chip, the amount of physical synapses can be reduced to 1/16 of that needed in the original network, while the accuracy is maintained. This compression technique can greatly reduce the number of SRAMs inside the chip as well as the power consumption of the chip. This design achieves throughput per unit area of 1.1 GSOP/( s·mm2) at 1.2 V, and energy consumed per SOP of 35 pJ. A 2-layer fully-connected spiking neural network is mapped to the chip, and thus the chip is able to realize handwritten digit recognition on MNIST with an accuracy of 91.2%.

中文翻译:

深度压缩技术在尖峰神经网络芯片中的应用。

在本文中,开发了包含192个神经元和6144个突触的可重构和可扩展的尖峰神经网络处理器。通过在尖峰神经网络芯片中使用深度压缩技术,可以将物理突触的数量减少到原始网络所需突触的数量的1/16,同时保持精度。这种压缩技术可以大大减少芯片内部的SRAM数量以及芯片的功耗。这种设计在1.2 V电压下的单位面积吞吐量为1.1 GSOP /(s·mm2),每SOP消耗的能量为35 pJ。2层全连接尖峰神经网络映射到该芯片,因此该芯片能够在MNIST上以91.2%的精度实现手写数字识别。
更新日期:2020-04-22
down
wechat
bug