FDS: Feedback-guided Domain Synthesis with Multi-Source Conditional Diffusion Models for Domain Generalization

TL;DR

FDS introduces a feedback-guided diffusion model approach to synthesize diverse pseudo-domains, improving model robustness against domain shifts by generating a broad spectrum of training data.

Contribution

The paper presents a novel diffusion-based method for domain synthesis that incorporates feedback to control diversity and span multiple distributions, enhancing domain generalization.

Findings

Sets new benchmarks in domain generalization performance

Effectively manages diverse types of domain shifts

Synthesizes broad distribution spectrum training data

Abstract

Domain Generalization techniques aim to enhance model robustness by simulating novel data distributions during training, typically through various augmentation or stylization strategies. However, these methods frequently suffer from limited control over the diversity of generated images and lack assurance that these images span distinct distributions. To address these challenges, we propose FDS, Feedback-guided Domain Synthesis, a novel strategy that employs diffusion models to synthesize novel, pseudo-domains by training a single model on all source domains and performing domain mixing based on learned features. By incorporating images that pose classification challenges to models trained on original samples, alongside the original dataset, we ensure the generation of a training set that spans a broad distribution spectrum. Our comprehensive evaluations demonstrate that this methodology sets new benchmarks in domain generalization performance across a range of challenging datasets, effectively managing diverse types of domain shifts. The code can be found at: \url{https://github.com/Mehrdad-Noori/FDS.git}.