设计一个集成模型,以时间图的注意力和变压器增强的RNN来增强异常检测的设计
Sai Babu Veesam1, Aravapalli Rama Satish1, Sreenivasulu Tupakula2
1School of Computer Science, VIT-AP University, Vijayawada, 522241, Andhra Pradesh, India.
Scientific reports
|January 21, 2025
概括
本研究引入了一种新的深度学习框架,用于监控系统中有效检测异常,通过增强时间和空间上下文建模,显著提高公共安全. 新方法实现了卓越的性能和通用性,用于检测看不见的异常.
科学领域:
- 计算机科学 计算机科学
- 人工智能的人工智能
- 机器学习 机器学习
背景情况:
- 目前监控中的异常检测方法在复杂的多摄像头环境中与长期的依赖性和空间相关性作斗争.
- 传统方法通常需要大量的标记数据,这限制了它们对新型,未见的异常进行概括的能力.
研究的目的:
- 开发一个先进的深度学习框架,用于在摄像头监控系统中稳健有效地检测异常.
- 改进时间依赖和空间相关性的建模,特别是在动态的多摄像头环境中.
- 增强异常检测系统对未见异常的概括能力,使用有限的标记数据.
主要方法:
- 循环神经网络 (RNN) 与图表注意网络 (GAT) 的集成,以增强时空建模.
- 利用转换器增强型RNN与自我注意机制,以改善时间上下文.
- 应用多模式变量自动编码器 (MVAE) 来融合各种传感器数据 (视频,音频,运动).
- 实施原型网络,用于少量学习,以解决有限的异常标签.
- 采用时空自编码器,以基于学习到的正常行为模式进行无监督的异常检测.
主要成果:
- 与传统模型相比,精度,回忆和F1分数 (10-15%) 显著改善.
- 证明了增强的概括能力,在新事件检测中获得了高达20%的收益.
- 多式联运数据的有效融合,显示了对噪声和缺失样本的弹性.
结论:
- 拟议的框架通过提高异常检测准确性和概括性,为现实世界监控系统提供了实质性的进步.
- 先进的深度学习技术的结合有效地解决了复杂环境中现有方法的局限性.
- 这种方法通过更可靠和更适应的异常检测能力来提高公共安全.
相关概念视频
01:27Steady Flow of a Fluid Stream
Consider a control volume, such as a pipe with solid boundaries, through which fluid flows and changes direction due to the impulse exerted by the resulting force from the pipe walls. In steady flow, the mass of fluid entering the control volume at a given time, t, with velocity v1, is equal to the mass leaving after infinitesimal time dt, with velocity v2.
During this process, the momentum of the fluid within the control volume remains constant over the time interval dt. By applying the...
During this process, the momentum of the fluid within the control volume remains constant over the time interval dt. By applying the...
01:14Conservation of Mass in Moving, Nondeforming Control Volume
Stormwater detention basins are essential in managing runoff during heavy rainfall, particularly in urban areas where impervious surfaces increase the risk of flooding. Understanding the conservation of mass in these systems allows engineers to optimize basin performance, balancing inflow, outflow, and water storage.
In the context of a detention basin, the conservation of mass states that the total mass of water entering the basin must equal the mass leaving the basin plus any accumulation of...
In the context of a detention basin, the conservation of mass states that the total mass of water entering the basin must equal the mass leaving the basin plus any accumulation of...
01:30Typical Model Studies
Fluid mechanics model studies often utilize scaled-down systems to predict fluid behavior in full-scale environments, such as river flows, dam spillways, and structures interacting with open surfaces. Maintaining Froude number similarity in river models is crucial, as it replicates surface flow features like wave patterns and velocities.
01:21Design Example: Creating a Hydraulic Model of a Dam Spillway
Scaled hydraulic models of dam spillways provide a practical way to replicate and study the intricate flow dynamics of these structures. Often built to a 1:15 ratio, these models allow for observing critical water behavior, such as velocity distribution, flow patterns, and energy dissipation.
01:29Gradually Varying Flow
Gradually varying flow (GVF) in open channels describes situations where water depth changes slowly along the channel due to factors like non-uniform bed slope, channel shape variations, or obstructions. This flow type occurs when the depth adjusts gradually to balance gravitational forces, shear forces, and energy requirements, resulting in a low rate of depth change.Characteristics of Gradually Varying FlowGVF is commonly observed in natural streams, rivers, and canals, where flow depth...
01:24Rapidly Varying Flow
Rapidly varying flow (RVF) in open channels is characterized by abrupt changes in flow depth over a short distance, with the rate of depth change relative to distance often approaching unity. These flows are inherently complex due to their transient and multi-dimensional nature, making exact analysis difficult. However, approximate solutions using simplified models provide valuable insights into their behavior.Key Features of Rapidly Varying FlowRVF is commonly observed in scenarios involving...