Search research articles

关于 JoVE

概览领导团队博客 JoVE 帮助中心

作者

出版流程编辑委员会范围与政策同行评审常见问题投稿

图书馆员

用户评价订阅访问资源图书馆顾问委员会常见问题

研究

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments 存档

教育

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual 教师资源中心教师网站

使用条款与条件

相关概念视频

Neural Circuits

Neural Circuits

Neural circuits and neuronal pools are two of the main structures found in the nervous system. Neural circuits are networks of neurons that work together to carry out a specific task or process. They consist of interconnected neurons and glial cells, which provide structural and metabolic support.
Neuronal pools are collections of nerve cells with similar functions and interact through chemical and electrical signals. These pools include both interneurons (the central neural circuit nodes that...

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序

Same author

ISTASTrack: Bridging ANN and SNN via ISTA Adapter for RGB-Event Tracking.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026

Same author

Subparticle Operando Imaging for Probing Electrocatalytic Intermediates and Cation Effects.

Journal of the American Chemical Society·2026

Same author

Adaptive dendritic plasticity in brain-inspired dynamic neural networks for enhanced multi-timescale feature extraction.

Neural networks : the official journal of the International Neural Network Society·2025

Same author

Adaptive spatiotemporal neural networks through complementary hybridization.

Nature communications·2024

Same author

Temporal dendritic heterogeneity incorporated with spiking neural networks for learning multi-timescale dynamics.

Nature communications·2024

Same author

Direct measurement of radial fluence distribution inside a femtosecond laser filament core.

Optics express·2020

Search research articles

相关实验视频

Updated: May 5, 2026

Examining Local Network Processing using Multi-contact Laminar Electrode Recording

Examining Local Network Processing using Multi-contact Laminar Electrode Recording

Published on: September 8, 2011

时间局部注意力与自适应解码:增强用于时间计算应用程序的尖端神经网络.

Hanxiao Fan¹, Hanle Zheng¹, Zikai Wang²

¹Center for Brain-Inspired Computing Research, Department of Precision Instrument, Tsinghua University, Beijing, 100084, China.

Neural networks : the official journal of the International Neural Network Society

|January 15, 2026

概括

尖端神经网络 (SNN) 难以处理长序列. 时间局部注意力 (TLA) 和自适应解码 (AD) 增强SNN的时间计算,提高性能和训练速度.

关键词:

适应性解码适应性解码序列学习的学习顺序.尖的神经网络的神经网络.时间局部关注时间局部关注

更多相关视频

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

Published on: March 25, 2014

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Published on: November 12, 2019

相关实验视频

Last Updated: May 5, 2026

Examining Local Network Processing using Multi-contact Laminar Electrode Recording

Examining Local Network Processing using Multi-contact Laminar Electrode Recording

Published on: September 8, 2011

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

Published on: March 25, 2014

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Using Neuron Spiking Activity to Trigger Closed-Loop Stimuli in Neurophysiological Experiments

Published on: November 12, 2019

科学领域:

人工智能的人工智能
计算神经科学是一种神经科学.

背景情况:

尖端神经网络 (SNN) 提供了复杂的学习和时间计算的潜力,因为它们受到大脑启发的动态和能源效率.
然而,SNN在长序渐变传播方面面临着挑战,这限制了它们在捕获时间特征方面的有效性.

研究的目的:

引入新的机制,以提高SNN在时间计算任务中的性能.
解决SNN训练和推理方面的局限性,特别是关于长序列和梯度传播.

主要方法:

引入时间局部注意 (TLA) 以减少序列长度和改善梯度流.
纳入自适应解码 (AD) 以基于性能相关性优化SNN输出.
整合TLA和AD (TLA-AD) 进行全面的SNN建模和培训.

主要成果:

TLA-AD方法显著提高了SNN的性能,并且在不增加参数数量的情况下加速了训练.
在DEAP数据集上实现了最先进的准确性 (93.52%的价值,93.41%的唤起).
与其他SNN方法相比,在SEED (87.41%) 和IMDB (86.31%) 数据集上表现出具有竞争力的准确性.

结论:

在时间计算中,TLA-AD为SNN提供了有效的优化策略.
提出的方法克服了SNN培训和绩效的关键局限性.
这项工作为在时间序列数据分析中更先进,更高效的SNN应用铺平了道路.