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

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

Identification and Validation of Reference Genes for Reliable RT-qPCR Normalization in <i>Schisandra chinensis</i> Across Different Tissues and Abiotic Stress Conditions.

Plants (Basel, Switzerland)·2026

Same author

Immediate implant placement in thin buccal bone phenotype: CBCT-based retrospective comparison between DBBM and H-PRF bone block.

Oral surgery, oral medicine, oral pathology and oral radiology·2026

Same author

The role of oxidative stress-induced ribosomal dysfunction in embryogenic decline of Schisandra chinensis.

Tree physiology·2026

Same author

Association Between Impaired Fasting Glucose and Adverse Outcomes in Patients Treated With Peritoneal Dialysis: A Retrospective Study From Southern China.

Endocrinology, diabetes & metabolism·2026

Same author

The association between asthma and ADHD in children and adolescents: An observational study and a meta-analysis of Mendelian randomization studies.

Medicine·2026

Same author

A review of nucleic acid aptamer-functionalized metal-organic frameworks for cancer biomarker detection.

Colloids and surfaces. B, Biointerfaces·2026

Search research articles

相关实验视频

Updated: Jul 11, 2025

Identifying Amino Acid Overproducers Using Rare-Codon-Rich Markers

Identifying Amino Acid Overproducers Using Rare-Codon-Rich Markers

Published on: June 24, 2019

GRAND:基于GAN的软件运行时异常检测方法,使用跟踪信息.

Shiyi Kong¹, Jun Ai², Minyan Lu²

¹School of Reliability and Systems Engineering, Beihang University, No. 37, Xueyuan Road, Haidian District, Beijing 100191, China; Beijing Institute of Astronautical Systems Engineering, No. 1, Nandahongmen Road, Fengtai District, Beijing 100076, China.

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

|November 4, 2023

概括

本研究介绍了GRAND,一种用于使用内部执行轨迹检测软件异常的神经网络. 它显著改善了卡桑德拉等复杂系统的故障检测,达到99%的F1得分.

关键词:

没有了,没有了,没有了.内部执行跟踪软件异常检测检测软件异常检测没有监督的学习学习.这就是VAE的意义.

更多相关视频

Utilization of Grafix for the Detection of Transient Interactors of Saccharomyces cerevisiae Spliceosome Subcomplexes

Utilization of Grafix for the Detection of Transient Interactors of Saccharomyces cerevisiae Spliceosome Subcomplexes

Published on: November 9, 2020

Workflow Based on the Combination of Isotopic Tracer Experiments to Investigate Microbial Metabolism of Multiple Nutrient Sources

Workflow Based on the Combination of Isotopic Tracer Experiments to Investigate Microbial Metabolism of Multiple Nutrient Sources

Published on: January 22, 2018

相关实验视频

Last Updated: Jul 11, 2025

Identifying Amino Acid Overproducers Using Rare-Codon-Rich Markers

Identifying Amino Acid Overproducers Using Rare-Codon-Rich Markers

Published on: June 24, 2019

Utilization of Grafix for the Detection of Transient Interactors of Saccharomyces cerevisiae Spliceosome Subcomplexes

Utilization of Grafix for the Detection of Transient Interactors of Saccharomyces cerevisiae Spliceosome Subcomplexes

Published on: November 9, 2020

Workflow Based on the Combination of Isotopic Tracer Experiments to Investigate Microbial Metabolism of Multiple Nutrient Sources

Workflow Based on the Combination of Isotopic Tracer Experiments to Investigate Microbial Metabolism of Multiple Nutrient Sources

Published on: January 22, 2018

科学领域:

计算机科学计算机科学
软件工程软件工程软件工程
人工智能的人工智能

背景情况:

软件运行时异常检测对于在关键故障之前识别系统故障至关重要.
现有的方法通常依赖于外部绩效指标,可能缺少内部功能错误.

研究的目的:

提出一种新的神经网络模型,GRAND,用于使用内部执行痕迹检测软件异常.
评估GRAND在识别性能问题以外的故障方面的有效性,包括功能错误.

主要方法:

开发了GRAND,一个神经网络,结合了变量自编码器和生成对抗网络.
利用来自卡桑德拉数据库系统的内部执行痕迹进行实证评估.
收集了一个大型数据集,包括5180个时间序列,每个数据点超过一千万个.

主要成果:

格兰达达到了99%的F1得分,表现优于基线模型 (93%和87%).
废除研究显示,通过确保任务一致性,工作负载信息提高了F1得分16%.
对数据融合的注意力机制提高了F1得分的32%.

结论:

格兰德有效地使用内部执行轨迹检测软件异常,超越现有方法.
该模型的架构,包括工作负载信息和注意力机制,有助于其卓越的性能.
这种方法提供了一种更全面的方法来确保复杂的软件系统的可靠性.