Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
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 JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关实验视频

Updated: Jan 8, 2026

Decoding Natural Behavior from Neuroethological Embedding
08:00

Decoding Natural Behavior from Neuroethological Embedding

Published on: October 3, 2025

556

通过对未标记的数据进行自我监督的预训,推进动物行为识别.

Axiu Mao1,2,3,4, Miaoyun Peng1, Guikun Liu1

  • 1School of Communication Engineering, Hangzhou Dianzi University, Hangzhou, 310018, Zhejiang Province, China.

Scientific reports
|December 16, 2025
PubMed
概括

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

PSMA-Targeted Radioligand Therapy Beyond the Post-Taxane Setting: A Review of Evidence Across the Prostate Cancer Spectrum.

Cancers·2026
Same author

Stable Agrobacterium-mediated transformation system for Flammulina filiformis using Geneticin (G418) as efficient selection agent.

Microbiological research·2026
Same author

Cerebral Foreign Body Granulomatous Reaction Secondary to Hydrophilic Polymer Embolization Post-Mechanical Thrombectomy: A Case Report and Overview of Differential Diagnosis.

Stroke (Hoboken, N.J.)·2026
Same author

HAdV55 reprograms host 3D genome architecture and mitochondrial metabolism to drive pathogenesis.

Cellular and molecular life sciences : CMLS·2026
Same author

Balancing Passivation and Conductivity: The Critical Role of Carbon Layer Thickness in Fe<sub>2</sub>O<sub>3</sub>/Fe<sub>3</sub>O<sub>4</sub> Heterojunction Photoanodes.

Langmuir : the ACS journal of surfaces and colloids·2026
Same author

Rapid and precise amplicon-based genome sequencing to preparedness and response chikungunya virus outbreaks.

Emerging microbes & infections·2026

本研究引入了使用跨物种未标记数据的自我监督学习框架,以改善动物活动识别 (AAR). 该方法通过有限的标记数据提高了性能,为可扩展的行为监控铺平了道路.

科学领域:

  • 计算机视觉 计算机视觉
  • 机器学习 机器学习
  • 动物行为分析 动物行为分析

背景情况:

  • 对于动物活动识别 (AAR) 的深度学习需要大型标记数据集.
  • 当前的预培训方法往往忽略了有价值的跨物种未标记的数据.

研究的目的:

  • 开发一个自我监督的学习框架,利用跨物种的未标记数据来解决AAR中的注释稀缺问题.
  • 提高动物行为监测的效率和可扩展性.

主要方法:

  • 一个两阶段的方法:自主监督的PatchTST编码器的预训练,使用跨物种未标记的数据,以时间频率一致性为目标.
  • 微调预先训练的编码器以一种新型的分类模型进行微调,该模型整合了本地和全球运动模式.

主要成果:

  • 在有限的标记数据条件下,在AAR中显著提高了性能.
  • 与基线相比,获得了4.79%的准确性和4.57%的F1得分改善.
  • 改善了对类似行为的歧视,并且在更少的样本中保持了稳定性.

结论:

  • 拟议的框架有效地利用跨物种未标记的数据,以获得标签效率高的AAR.
关键词:
动物行为 动物行为自主监督学习学习时间序列时间序列时间频率一致性时间频率一致性变压器模型模型

更多相关视频

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
05:41

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

9.8K
Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction
06:19

Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction

Published on: August 16, 2024

795

相关实验视频

Last Updated: Jan 8, 2026

Decoding Natural Behavior from Neuroethological Embedding
08:00

Decoding Natural Behavior from Neuroethological Embedding

Published on: October 3, 2025

556
A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis
05:41

A Step-by-Step Implementation of DeepBehavior, Deep Learning Toolbox for Automated Behavior Analysis

Published on: February 6, 2020

9.8K
Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction
06:19

Integration of Animal Behavioral Assessment and Convolutional Neural Network to Study Wasabi-Alcohol Taste-Smell Interaction

Published on: August 16, 2024

795
  • 建立了可扩展和数据效率高的动物行为监测系统的新方向.