Scalable and Interpretable One-class SVMs with Deep Learning and Random Fourier features

TL;DR

This paper introduces AE-1SVM, a novel end-to-end deep learning approach combining autoencoders and random Fourier features to improve large-scale, high-dimensional anomaly detection with enhanced interpretability.

Contribution

It presents the first integrated deep learning framework for one-class SVMs that jointly learns representations and decision boundaries, enabling better scalability and interpretability.

Findings

Significantly outperforms previous separate training methods.

Enables gradient-based interpretability for anomaly detection.

Effective on diverse unsupervised anomaly detection tasks.

Abstract

One-class support vector machine (OC-SVM) for a long time has been one of the most effective anomaly detection methods and extensively adopted in both research as well as industrial applications. The biggest issue for OC-SVM is yet the capability to operate with large and high-dimensional datasets due to optimization complexity. Those problems might be mitigated via dimensionality reduction techniques such as manifold learning or autoencoder. However, previous work often treats representation learning and anomaly prediction separately. In this paper, we propose autoencoder based one-class support vector machine (AE-1SVM) that brings OC-SVM, with the aid of random Fourier features to approximate the radial basis kernel, into deep learning context by combining it with a representation learning architecture and jointly exploit stochastic gradient descent to obtain end-to-end training. Interestingly, this also opens up the possible use of gradient-based attribution methods to explain the decision making for anomaly detection, which has ever been challenging as a result of the implicit mappings between the input space and the kernel space. To the best of our knowledge, this is the first work to study the interpretability of deep learning in anomaly detection. We evaluate our method on a wide range of unsupervised anomaly detection tasks in which our end-to-end training architecture achieves a performance significantly better than the previous work using separate training.