为了丰富当前GAD数据集的种类并克服其局限性,我们提出了DGraph,这是一种真实的大规模动态图,由超过3M个节点和4M个边组成。DGraph由Finvolution Group提供。它代表了金融行业中的真实社会网络。一个节点代表一个Finvolution用户,从一个用户到另一个用户的边缘意味着用户将另一个用户视为紧急联系人。此外,DGraph中的异常...
DGraph: A Large-Scale Financial Dataset for Graph Anomaly Detection DGragh是一个用于图异常检测(gragh anomaly detection, GAD)的大型金融数据集。 它包含300w个节点、400w个动态边和100w个ground-truth节点。 作者在关注的问题 大领域:图异常检测(GAD) 现实世界场景中,异常是普遍存在的且具有破坏性的。作者举...
概述 信也科技发布了一个名为DGraph的大型金融数据集，旨在支持金融领域，尤其是反欺诈研究。DGraph的独特之处在于提出了背景节点的概念，强调在真实场景中，不需对所有节点进行分类或预测，以避免破坏网络结构的连通性和丢失丰富信息。然而，当前许多GAD数据集忽视了背景节点的存在。该数据集由Finvolution Gro...
Flow-based anomaly detection is gaining momentum because it can be deployed for real time detection as it analyses only packet headers. To evaluate anomaly detection techniques, labeled dataset is required as unlabeled dataset is not useful for the evaluation. Many packet based network traffic dataset...
An Anomaly Intrusion Detection Approach Using Cellular Neural Networks. This paper presents an anomaly detection approach for the network intrusion detection based on Cellular Neural Networks (CNN) model. CNN has features with ... Z Yang,A Karahoca - International Conference on Computer & Information...
AD3: Introducing a score for Anomaly Detection Dataset Difficulty assessment using VIADUCT dataset 德基先生 拉黑单纯懒得和二极管浪费时间,不如敲代码有意思摘要: 近年来,视觉工业异常检测(IAD)领域涌现出许多新的半监督学习方法。同时,用于基准测试这些方法的新数据集却寥寥无几。最受欢迎的数据集是MVTec-AD数据...
Faster R-CNN Concrete Crack Detection Multi-scale Convolutional Denoising Autoencoder Network Model CNN for Classfication Weibull Neuro-Evolution GAN for Defect Detection GAN for Anomaly Detection GAN for Defect Classfication YOLO for Defect Classfication ...
- Dowload "C01_create_small_INT_dataset", "E01_simple_AE_test", and "anomaly_conditions" - Run "make_dataset_for_car_and_conveyor.py" and "make_dataset_for_train.py" in "C01_create_small_INT_dataset" to make dataset. [20 Feb. 2020] Note that the description of the gain parameters...
MVTec 3D anomaly detection dataset (MVTec 3D-AD)is a comprehensive 3D dataset for the task of unsupervised anomaly detection and localization. It contains over 4000 high-resolution scans acquired by an industrial 3D sensor. Each of the 10 different object categories comprises a set of defect-fre...
问题分了视频理解和视频推理两大类,涵盖了包括异常检测、交互理解、逻辑推理和因果推理在内的各种task,目前VLM比较薄弱的包括:temporal location, object tracking, and anomaly detection。14个视频类别:Howto & Style(H&S),新闻与政治(N&P),宠物与动物(P&A),汽车与车辆(A&E),游戏(Gam.),电影与动画(F&A),...