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

Search research articles

相关实验视频

Updated: Jan 8, 2026

Trajectory Data Analyses for Pedestrian Space-time Activity Study

Trajectory Data Analyses for Pedestrian Space-time Activity Study

Published on: February 25, 2013

ADP-Net:一个分层的注意力-扩散-预测框架,用于预测人类的轨迹.

Zhenggui Zhang¹, Shanlin Xiao¹, Zhiyi Yu¹

¹School of Microelectronics Science and Technology, Sun Yat-sen University, Guangzhou, China.

Frontiers in artificial intelligence

|December 15, 2025

概括

对于自主系统而言,精确的群众行为预测通过注意力扩散预测网络 (ADP-Net) 得到了改进. 这种新的框架增强了时空建模,在预测准确性方面超过了现有的方法.

相关概念视频

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

SHP2 Inhibition Reveals Compensatory PI3K-AKT Activation in KRAS-Driven Pancreatic Cancer: Discovery of <b>SDUY104</b> and Rational Approaches for Combination Therapy.

Journal of medicinal chemistry·2026

Same author

miR-27a Suppresses Mitochondrial Function to Promote Hepatic Steatosis in High-Fat-Diet-Induced Obesity.

Molecules (Basel, Switzerland)·2026

Same author

SHP2-targeted cancer therapeutics: Structural insights, inhibitor development and emerging modalities.

Bioorganic & medicinal chemistry·2026

Same author

Dual Allosteric Targeting of SHP2: Development of Bivalent Inhibitors with Enhanced Potency for Cancer Treatment.

Journal of medicinal chemistry·2026

Same author

Variable effects of biochar on soil greenhouse gas emissions: A meta-analysis of climate, soil, and biochar property interactions.

Environmental research·2026

Same author

CO<sub>2</sub>-sensitive K<sup>+</sup> channel traffic affects stomata and whole-plant water use.

Journal of integrative plant biology·2026

科学领域:

计算机视觉计算机视觉
人工智能的人工智能
机器人技术机器人技术机器人技术

背景情况:

人群行为预测对自主系统至关重要.
当前的时空图卷积网络 (STGCNs) 存在局限性,包括受限的受容场和差的拓扩展性.
这些局限性导致模拟复杂人群动态时的性能不足.

研究的目的:

为准确预测人群行为开发一种新的计算框架.
解决现有的STGCN在捕捉时空动态方面的局限性.
建议注意力扩散预测网络 (ADP-Net) 用于增强人群行为建模.

主要方法:

建立了图形卷积网络 (GCN) 和个性化传播神经架构之间的理论联系.
引入了注意力扩散预测网络 (ADP-Net),其中包括一致的图形卷积层,多尺度的注意力扩散层 (实施图形扩散卷积 - GDC) 和自适应时间卷积模块.
采用GCN操作的多项式近似和GDC的近似个性化传播方案,以实现高效的多跳互动建模.

主要成果:

在标准基准 (ETH/UCY和斯坦福无人机数据集) 上取得了最先进的结果.
与先前的方法相比,已经证明了显著的改善,平均位移误差 (ADE) 提高了4%,最终位移误差 (FDE) 提高了26%.
验证了框架在建模时空人群动态方面的有效性.

结论:

拟议的ADP-Net为在自主系统中预测人群行为提供了一个强大的理论框架和实际实施.
ADP-Net克服了传统的STGCN的局限性,通过实现高效的多节点交互建模和保持结构一致性.
这一进步有助于在人类环境中实现更安全,更有效的自主系统运行.

关键词:

个性化的PageRank 个人化图形的扩散卷积曲线.图形神经网络的神经网络多跳跃式多跳跃式代表性学习学习学习时空关系建模时空关系建模轨迹预测轨迹预测

更多相关视频

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024

相关实验视频

Last Updated: Jan 8, 2026

Trajectory Data Analyses for Pedestrian Space-time Activity Study

Trajectory Data Analyses for Pedestrian Space-time Activity Study

Published on: February 25, 2013

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Swin-PSAxialNet: An Efficient Multi-Organ Segmentation Technique

Published on: July 5, 2024