异常检测已经得到了广泛的研究和应用。建立一个有效的异常检测系统需要研究者和开发者从嘈杂的数据中学习复杂的结构,识别动态异常模式,用有限的标签检测异常。与经典方法相比,近年来深度学习技术的进步极大地提高了异常检测的性能,并将异常检测扩展到广泛的应用领域。本教程将帮助读者全面理解各种应用领域中基于深度学习的异常检测技术。首先,我们概述了异常检测问题,介绍了在深度模型时代之前采用的方法,并列出了它们所面临的挑战。然后我们调查了最先进的深度学习模型,范围从构建块神经网络结构,如MLP, CNN,和LSTM,到更复杂的结构,如自动编码器,生成模型(VAE, GAN,基于流的模型),到深度单类检测模型,等等。此外,我们举例说明了迁移学习和强化学习等技术如何在异常检测问题中改善标签稀疏性问题,以及在实际中如何收集和充分利用用户标签。其次,我们讨论来自LinkedIn内外的真实世界用例。本教程最后讨论了未来的趋势。
https://sites.google.com/view/kdd2020deepeye/home
Part 1. Introduction (30 min)
1.1. Overview of Anomaly Detection
1.2. Anomaly Detection Application and Challenges
1.3. Traditional Techniques and Motivation for Deep Learning
Part 2. Deep Learning for Anomaly Detection (90 min)
a. Integrated Semi-Supervised Learning
b. Data Augmentation and Transfer Learning
a. Deep One-Class Models (Deep OC)
b. AutoEncoder (AE)
c. Variational AutoEncoder (VAE)
d. Generative Adversarial Networks (VAE, GAN, Flow-based)
a. MultiLayer Perceptron (MLP)
b. Convolutional Neural Networks (CNN)
c. Recurrent Neural Networks (RNN)
2.1. Basic Building Blocks
2.2 Fundamental Model Structures Applied to Anomaly Detection Tasks
2.3. Compensate for Sparse Labels
Part 3. Real-world Applications for Anomaly Detection (50 min)
Algorithms and Evaluation
System Architecture
Usability in Production
3.1 Anomaly Detection for Autonomous Vehicle Development
3.2 Anomaly Detection at LinkedIn
Part 4. Conclusion and Future Trends (10 min)
专知便捷查看
便捷下载,请关注专知公众号(点击上方蓝色专知关注)
后台回复“DLAD” 可以获取《【KDD2020-Tutorial】深度学习异常检测,180页ppt》专知下载链接索引