---
格式版本: 2
标题: "Noisy group neurons with synchronous resetting for high-performance spiking neural networks"
原文链接: "https://arxiv.org/abs/2608.17394"
发布日期: "2026-08-18"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Tue, 18 Aug 2026 05:42:11 UTC (5,334 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-20T15:58:47+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-20T15:54:04+08:00"
入库时间: "2026-08-20T07:58:47.175Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=inference%20networking&searchtype=all"
匹配关键词:
  - "performance"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 10
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "该arXiv页面为脉冲神经网络（SNN）学术论文，内容涉及神经形态计算，与超节点、AI Rack、机柜级AI基础设施、供电散热互连等主题完全无关，且无任何厂商或商业信号。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-20T16:03:51+08:00"
AI主题相关性: 0
AI来源权威性: 5
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 5
AI摘要: "研究人员提出带群体同步重置和神经随机性的噪声群神经元（NGN）模型，结合基于平均场动力学的反向传播训练脉冲神经网络，在CIFAR等多项基准上验证有效性，并在CIFAR10-DVS上以10个推理时间步达到87.35%的准确率。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:22:27.192Z"
采集批次: "2026年8月20日14点19分32秒"
采集批次ID: "20260820-141932-079"
去重键: "https://arxiv.org/abs/2608.17394"
---

## Computer Science > Computer Vision and Pattern Recognition

## Title:Noisy group neurons with synchronous resetting for high-performance spiking neural networks

Authors:[Yajie Zhai](https://arxiv.org/search/cs?searchtype=author&query=Zhai,+Y), [Yanmei Kang](https://arxiv.org/search/cs?searchtype=author&query=Kang,+Y), [Meng Li](https://arxiv.org/search/cs?searchtype=author&query=Li,+M), [Zigang Huang](https://arxiv.org/search/cs?searchtype=author&query=Huang,+Z)

[View PDF](https://arxiv.org/pdf/2608.17394) [HTML (experimental)](https://arxiv.org/html/2608.17394v1)

> Abstract:Spiking neural networks (SNNs), characterized by bio-inspired neuronal dynamics and event-driven communication, have attained significant progress in recent years. Nevertheless, training deep SNNs remains challenging due to spatiotemporal information loss and gradient mismatching. To simultaneously address these issues, we propose a noisy group neuron (NGN) model, which incorporates population-level synchronous resetting and neural stochasticity as fundamental computational mechanisms. We then develop the NGN method as a framework that combines the NGN model with backpropagation learning based on mean-field dynamics. We demonstrate the advantages of the NGN method through theoretical analysis and experimental validation on CIFAR-10, CIFAR-100, Tiny-ImageNet, DVS-Gesture, N-Caltech101, and CIFAR10-DVS. The proposed approach achieves an accuracy of 87.35% on CIFAR10-DVS within 10 inference time steps. These results support NGN as a practical approach to high-performance neuromorphic computing.

| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| --- | --- |
| Cite as: | [arXiv:2608.17394](https://arxiv.org/abs/2608.17394) \[cs.CV\] |
|  | (or [arXiv:2608.17394v1](https://arxiv.org/abs/2608.17394v1) \[cs.CV\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.17394](https://doi.org/10.48550/arXiv.2608.17394) |

## Submission history

From: Yajie Zhai \[[view email](https://arxiv.org/show-email/bcb8d6ca/2608.17394)\]  
**\[v1\]** Tue, 18 Aug 2026 05:42:11 UTC (5,334 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.17394) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))