基于模拟非易失性存储器的神经形态架构的器件与系统级设计考量
神经与进化计算
2016-05-09 v2 人工智能
摘要
本文概述了脑启发计算领域的近期进展，重点聚焦于使用新兴存储器作为电子突触的实现方式。给出了设计考量与挑战，例如关于多电平状态、器件差异性、编程能耗、阵列级连接性、扇入/扇出、连线能耗以及 IR 压降的需求与设计目标。连线在设计决策中日益重要，尤其对于大型系统，且周期间差异对学习性能有很大影响。
引用
@article{arxiv.1512.08030,
title = {Device and System Level Design Considerations for Analog-Non-Volatile-Memory Based Neuromorphic Architectures},
author = {Sukru Burc Eryilmaz and Duygu Kuzum and Shimeng Yu and H. -S. Philip Wong},
journal= {arXiv preprint arXiv:1512.08030},
year = {2016}
}
备注
4 pages, In Electron Devices Meeting (IEDM), 2015 IEEE International (pp. 4.1). IEEE. Original paper can be found here: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7409622. Abstract can be found here: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=7409622&refinements%3D4224410500%26filter%3DAND%28p_IS_Number%3A7409598%29