New Perspective on Progressive GANs Distillation for One-class Novelty Detection

TL;DR

This paper introduces a novel GAN-based approach for one-class novelty detection using latent space distances, and proposes Progressive Knowledge Distillation to enhance model efficiency and performance.

Contribution

It presents a new EDE-GAN architecture for anomaly detection and a progressive distillation method to create compact, high-performing GAN models.

Findings

EDE-GAN achieves state-of-the-art anomaly detection performance.

Progressive Knowledge Distillation improves GAN performance significantly.

Model compression ratios up to 700:1 with minimal performance loss.

Abstract

One-class novelty detection is conducted to identify anomalous instances, with different distributions from the expected normal instances. In this paper, the Generative Adversarial Network based on the Encoder-Decoder-Encoder scheme (EDE-GAN) achieves state-of-the-art performance. The two factors bellow serve the above purpose: 1) The EDE-GAN calculates the distance between two latent vectors as the anomaly score, which is unlike the previous methods by utilizing the reconstruction error between images. 2) The model obtains best results when the batch size is set to 1. To illustrate their superiority, we design a new GAN architecture, and compare performances according to different batch sizes. Moreover, with experimentation leads to discovery, our result implies there is also evidence of just how beneficial constraint on the latent space are when engaging in model training. In an attempt to learn compact and fast models, we present a new technology, Progressive Knowledge Distillation with GANs (P-KDGAN), which connects two standard GANs through the designed distillation loss. Two-step progressive learning continuously augments the performance of student GANs with improved results over single-step approach. Our experimental results on CIFAR-10, MNIST, and FMNIST datasets illustrate that P-KDGAN improves the performance of the student GAN by 2.44%, 1.77%, and 1.73% when compressing the computationat ratios of 24.45:1, 311.11:1, and 700:1, respectively.