用磁隧道结实现可生成正负脉冲的神经元，提升存算一体效率。
Signed Spiking Neuron Enabled by an Orthogonal-Easy-Axis Magnetic Tunnel Junction
- 通过正交易轴结构设计，让磁性器件模拟带符号的脉冲神经网络行为。
- 在10nm×45nm×50nm器件上实现91.06%准确率（CIFAR-10），接近理想模型性能。
- 适合用于低功耗类脑芯片，尤其面向事件驱动型视觉任务。
带符号脉冲神经元比标准脉冲神经元携带更丰富的信息。本文提出一种基于磁隧道结（MTJ）的紧凑型神经元，支持带符号漏积分放电（signed LIF）操作。利用自由层与钉扎层的正交易轴结构，该器件可实现双向脉冲输出，并将磁矩动态映射为带符号的LIF膜电位演化过程。朗道-利夫希茨-吉尔伯特模拟表明，合理设计自由层尺寸可使器件响应符合带符号LIF方程。一个典型尺寸为10 nm × 45 nm × 50 nm的器件，长宽高比例约为2:9:10。基于拟合器件神经元模型的网络评估在CIFAR-10上达到91.06%准确率，在CIFAR10-DVS上达到77.40%，基本保留了理想带符号LIF神经元的精度。
原文摘要 · Abstract (English)
Signed spiking neurons carry richer information than standard spiking neurons. This work proposes a compact magnetic tunnel junction (MTJ)-based neuron for signed leaky integrate-and-fire (LIF) operation. With orthogonal easy axes in the free and pinned layers, the device enables bipolar spike generation and maps magnetic-moment dynamics to signed LIF membrane-potential evolution. Landau--Lifshitz--Gilbert simulations show that proper free-layer dimensions allow the device response to follow a signed LIF equation. A representative design of 10 nm x 45 nm x 50 nm corresponds to an aspect ratio of about 2:9:10. Network evaluations using the fitted device-neuron model achieve 91.06% on CIFAR-10 and 77.40% on CIFAR10-DVS, retaining most of the accuracy of ideal signed LIF neurons.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。