在数据挖掘中,异常检测(英语:anomaly detection)对不符合预期模式或数据集中其他项目的项目、事件或观测值的识别。通常异常项目会转变成银行欺诈、结构缺陷、医疗问题、文本错误等类型的问题。异常也被称为离群值、新奇、噪声、偏差和例外。 特别是在检测滥用与网络入侵时,有趣性对象往往不是罕见对象,但却是超出预料的突发活动。这种模式不遵循通常统计定义中把异常点看作是罕见对象,于是许多异常检测方法(特别是无监督的方法)将对此类数据失效,除非进行了合适的聚集。相反,聚类分析算法可能可以检测出这些模式形成的微聚类。 有三大类异常检测方法。[1] 在假设数据集中大多数实例都是正常的前提下,无监督异常检测方法能通过寻找与其他数据最不匹配的实例来检测出未标记测试数据的异常。监督式异常检测方法需要一个已经被标记“正常”与“异常”的数据集,并涉及到训练分类器(与许多其他的统计分类问题的关键区别是异常检测的内在不均衡性)。半监督式异常检测方法根据一个给定的正常训练数据集创建一个表示正常行为的模型,然后检测由学习模型生成的测试实例的可能性。

VIP内容

检测语义异常具有挑战性,因为它们可能以无数种方式出现在真实数据中。虽然增强网络的健壮性可能足以对简单异常进行建模,但没有一种已知的好方法来准备模型,以应对可能发生的所有潜在的、看不见的异常,比如出现新的对象类。在本文中,我们展示了一个以前被忽视的异常检测(AD)策略,即对从一些大型和多样的语义任务转移过来的表示引入一个显式的归纳偏差。我们在使用干预的对照试验中严格验证了我们的假设,并表明它产生了出乎意料的有效辅助目标,优于以往的AD范式。

http://proceedings.mlr.press/v139/deecke21a.html

成为VIP会员查看完整内容
0
9

最新内容

Industrial processes are monitored by a large number of various sensors that produce time-series data. Deep Learning offers a possibility to create anomaly detection methods that can aid in preventing malfunctions and increasing efficiency. But creating such a solution can be a complicated task, with factors such as inference speed, amount of available data, number of sensors, and many more, influencing the feasibility of such implementation. We introduce the DeTAVIZ interface, which is a web browser based visualization tool for quick exploration and assessment of feasibility of DL based anomaly detection in a given problem. Provided with a pool of pretrained models and simulation results, DeTAVIZ allows the user to easily and quickly iterate through multiple post processing options and compare different models, and allows for manual optimisation towards a chosen metric.

0
0
下载
预览

最新论文

Industrial processes are monitored by a large number of various sensors that produce time-series data. Deep Learning offers a possibility to create anomaly detection methods that can aid in preventing malfunctions and increasing efficiency. But creating such a solution can be a complicated task, with factors such as inference speed, amount of available data, number of sensors, and many more, influencing the feasibility of such implementation. We introduce the DeTAVIZ interface, which is a web browser based visualization tool for quick exploration and assessment of feasibility of DL based anomaly detection in a given problem. Provided with a pool of pretrained models and simulation results, DeTAVIZ allows the user to easily and quickly iterate through multiple post processing options and compare different models, and allows for manual optimisation towards a chosen metric.

0
0
下载
预览
Top