Semi-supervised learning and inference in domain-wall magnetic tunnel junction (DW-MTJ) neural networks

Semi-supervised learning and inference in domain-wall magnetic tunnel junction (DW-MTJ) neural networks
复制标题

磁畴壁磁隧道结 (DW-MTJ) 神经网络中的半监督学习和推理

DOI:
10.1117/12.2530308
复制
发表时间:
2019
期刊:
SPIE Spintronics XII
影响因子:
--
通讯作者:
Marinella, Matthew M.
Marinella, Matthew M.
中科院分区:
--
文献类型:
--
作者:
Bennett, Christopher H.;Hassan, Naimul;Hu, Xuan;Incornvia, Jean Anne;Friedman, Joseph S.;Marinella, Matthew M.

文献摘要

参考文献

被引文献

相似文献

机器智能的进步已经引发了对实现这些算法的硬件加速器的兴趣,然而嵌入式电子设备具有严格的功率、面积预算和速度要求,这可能限制非易失性存储器(NVM)集成。在这种情况下,使用最少的训练数据的快速纳米磁神经网络的发展是有吸引力的。在这里,我们扩展了一个仅推理的建议,使用域壁MTJ(DW-MTJ)神经元的内在物理学进行在线学习,以实现完全无监督的模式识别操作,使用包含随机或塑性突触(权重)的赢家通吃网络。同时,读出层以监督的方式训练。我们发现,相对于竞争对手的忆阻神经网络提案,我们提出的设计可以在任务上接近最先进的成功,同时消除了使用CMOS器件构建神经元层通常所需的大部分面积和能量开销。
Advances in machine intelligence have sparked interest in hardware accelerators to implement these algorithms, yet embedded electronics have stringent power, area budgets, and speed requirements that may limit non- volatile memory (NVM) integration. In this context, the development of fast nanomagnetic neural networks using minimal training data is attractive. Here, we extend an inference-only proposal using the intrinsic physics of domain-wall MTJ (DW-MTJ) neurons for online learning to implement fully unsupervised pattern recognition operation, using winner-take-all networks that contain either random or plastic synapses (weights). Meanwhile, a read-out layer trains in a supervised fashion. We find our proposed design can approach state-of-the-art success on the task relative to competing memristive neural network proposals, while eliminating much of the area and energy overhead that would typically be required to build the neuronal layers with CMOS devices.
DOI: 10.1145/3309880
发表时间: 2019-03
期刊: ACM Journal on Emerging Technologies in Computing Systems (JETC)
影响因子: --
作者:
Ankit Mondal;Ankur Srivastava-
通讯作者: Ankit Mondal;Ankur Srivastava-
具有纳米级忆阻设备的仿生网络,结合了无监督和监督学习方法
DOI: 10.1145/2765491.2765528
发表时间: 2012
期刊: 2012 IEEE/ACM International Symposium on Nanoscale Architectures (NANOARCH)
影响因子: --
作者:
D. Querlioz;Weisheng Zhao;P. Dollfus;Jacques;O. Bichler;C. Gamrat
通讯作者: C. Gamrat
DOI: 10.1038/srep31932
发表时间: 2016-09-07
期刊: Scientific reports
影响因子: 4.6
作者:
Lin YP;Bennett CH;Cabaret T;Vodenicarevic D;Chabi D;Querlioz D;Jousselme B;Derycke V;Klein JO
通讯作者: Klein JO