• Financial System
  • OA
  • Bidding
  • Email
  • 简体中文
  • About Us
    • Overview
    • History
    • Organization
    • Our Team
      • Management
      • Research & Development
      • Functional Personnel
  • Research Divisions
    • Division of Fundamental Research
    • Center for General Software and Technologies of Big Data
      • Optimization Solver Development Laboratory
    • Center for Big Data Applications and Technologies
      • Data-driven Intelligent Information System Laboratory
      • Medical Big Data Laboratory
      • Human Language Technology Laboratory
      • Laboratory for Smart City, Transportation and Logistics Big Data
      • Laboratory for Public and Commercial Big Data
      • Shenzhen International Center for Industrial and Applied Mathematics (SICIAM) 
  • Our Research
    • R&D Projects
    • Educational Programs
      • CUHKSZ-SRIBD Joint PhD&Postdoc Programs
      • Scholarship Program
    • Seminars & Conferences
  • News
    • SRIBD News
    • MIIS
    • Audios & Videos
      • Videos
      • Periodicals
  • Recruiting
    • Reserach Scientist
    • Engineer
    • Administration Staff
    • Recruitment Overview
  • Contact Us
    • Inquiry
    • Public Relations
    • Cooperation
  • About Us
    • Overview
    • History
    • Organization
    • Our Team
      • Management
      • Research & Development
      • Functional Personnel
  • Research Divisions
    • Division of Fundamental Research
    • Center for General Software and Technologies of Big Data
      • Optimization Solver Development Laboratory
    • Center for Big Data Applications and Technologies
      • Data-driven Intelligent Information System Laboratory
      • Medical Big Data Laboratory
      • Human Language Technology Laboratory
      • Laboratory for Smart City, Transportation and Logistics Big Data
      • Laboratory for Public and Commercial Big Data
      • Shenzhen International Center for Industrial and Applied Mathematics (SICIAM) 
  • Our Research
    • R&D Projects
    • Educational Programs
      • CUHKSZ-SRIBD Joint PhD&Postdoc Programs
      • Scholarship Program
    • Seminars & Conferences
  • News
    • SRIBD News
    • MIIS
    • Audios & Videos
      • Videos
      • Periodicals
  • Recruiting
    • Reserach Scientist
    • Engineer
    • Administration Staff
    • Recruitment Overview
  • Contact Us
    • Inquiry
    • Public Relations
    • Cooperation
  • Financial System
  • OA
  • Bidding
  • Email
  • 简体中文
  • About Us
    • Overview
    • History
    • Organization
    • Our Team
      • Management
      • Research & Development
      • Functional Personnel
  • Research Divisions
    • Division of Fundamental Research
    • Center for General Software and Technologies of Big Data
      • Optimization Solver Development Laboratory
    • Center for Big Data Applications and Technologies
      • Data-driven Intelligent Information System Laboratory
      • Medical Big Data Laboratory
      • Human Language Technology Laboratory
      • Laboratory for Smart City, Transportation and Logistics Big Data
      • Laboratory for Public and Commercial Big Data
      • Shenzhen International Center for Industrial and Applied Mathematics (SICIAM) 
  • Our Research
    • R&D Projects
    • Educational Programs
      • CUHKSZ-SRIBD Joint PhD&Postdoc Programs
      • Scholarship Program
    • Seminars & Conferences
  • News
    • SRIBD News
    • MIIS
    • Audios & Videos
      • Videos
      • Periodicals
  • Recruiting
    • Reserach Scientist
    • Engineer
    • Administration Staff
    • Recruitment Overview
  • Contact Us
    • Inquiry
    • Public Relations
    • Cooperation

Breadcrumb

  • Home
  • Our Research
  • R&D Projects
  • Perturbation Learning Based Anomaly Detection

Perturbation Learning Based Anomaly Detection

Mar 28,2023 Projects

Project description/goals

Anomaly detection (AD) is an important research problem in many areas such as computer vision. AD aims to identify abnormal data from normal data and is usually an unsupervised learning task because the anomaly samples are unknown in the training stage.  

Importance/impact, challenges/pain points

Classical AD methods such as OCSVM [Schölkopf et al., 2001] and DSVDD [Ruff et al., 2018] require specific assumptions (e.g. hypersphere) for the distribution or structure of the normal data. The GAN-based approaches [Deecke et al., 2018; Perera et al., 2019] suffer from the instability problem of min-max optimization and have high computational costs.

Solution description

We propose a novel AD method called perturbation learning based anomaly detection (PLAD). PLAD aims to learn a perturbator and a classifier from the normal training data. The perturbator uses minimum effort to perturb the normal data to abnormal data while the classifier is able to classify the normal data and perturbed data into two classes correctly.

Key contribution/commercial implication

Compared with the state-of-the-art of anomaly detection, our method does not require any assumption about the shape (e.g. hypersphere) of the decision boundary and has fewer hyper-parameters to determine. Empirical studies on benchmark datasets verify the effectiveness and superiority of our method.

Next steps

Adapt the method to graph data and time series. Provide more theoretical analysis.

Team/contributors

Jinyu Cai (Fuzhou University), Visiting Student at SRIBD

Jicong Fan, SRIBD & CUHK-Shenzhen

 
Follow Us
  •  
  •  
  •  
  •  
  • Contact Us
    • Inquiry
    • Public Relations
    • Cooperation
  • Collaborative Institutions
    • The Chinese University of Hong Kong, Shenzhen
    • National Health Data Institute (Shenzhen)
    • Shenzhen International Center for Industrial and Applied Mathematics

  • Friendly Links
Copyright 2022 All Rights Reserved | 粤ICP备16049670号