Multi-Scale Architectures Matter: On the Adversarial Robustness of Flow-based Lossless Compression

TL;DR

This paper investigates how multi-scale architectures in flow-based models influence their adversarial robustness and effectiveness in lossless compression, highlighting the importance of architectural choices for performance and efficiency.

Contribution

The study emphasizes the significance of multi-scale architecture in flow-based models for enhancing adversarial robustness and compression performance.

Findings

Multi-scale architecture improves robustness against adversarial attacks.

Flow-based models with multi-scale design achieve better compression efficiency.

Multi-scale structures reduce computational complexity in flow models.

Abstract

As a probabilistic modeling technique, the flow-based model has demonstrated remarkable potential in the field of lossless compression \cite{idf,idf++,lbb,ivpf,iflow},. Compared with other deep generative models (eg. Autoregressive, VAEs) \cite{bitswap,hilloc,pixelcnn++,pixelsnail} that explicitly model the data distribution probabilities, flow-based models perform better due to their excellent probability density estimation and satisfactory inference speed. In flow-based models, multi-scale architecture provides a shortcut from the shallow layer to the output layer, which significantly reduces the computational complexity and avoid performance degradation when adding more layers. This is essential for constructing an advanced flow-based learnable bijective mapping. Furthermore, the lightweight requirement of the model design in practical compression tasks suggests that flows with multi-scale architecture achieve the best trade-off between coding complexity and compression efficiency.