SRasP: Self-Reorientation Adversarial Style Perturbation for Cross-Domain Few-Shot Learning

2026-03-05

AI Summary

This work addresses the limitations in cross-domain few-shot learning caused by domain shift due to style discrepancies and the instability of existing adversarial style perturbation methods, which often suffer from gradient instability and convergence to sharp minima that hinder generalization. To overcome these issues, the authors propose a self-redirected adversarial style perturbation approach that leverages global semantic guidance to identify inconsistent image regions, integrates local and global style gradients, and employs a multi-objective optimization function to enhance visual diversity while preserving semantic consistency. This method achieves a more stable, flat, and transferable solution space, significantly outperforming state-of-the-art approaches across multiple cross-domain few-shot benchmarks and demonstrating superior generalization and robustness.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Cross-Domain Few-Shot Learning (CD-FSL) aims to transfer knowledge from a seen source domain to unseen target domains, serving as a key benchmark for evaluating the robustness and transferability of models. Existing style-based perturbation methods mitigate domain shift but often suffer from gradient instability and convergence to sharp minima.To address these limitations, we propose a novel crop-global style perturbation network, termed Self-Reorientation Adversarial \underline{S}tyle \underline{P}erturbation (SRasP). Specifically, SRasP leverages global semantic guidance to identify incoherent crops, followed by reorienting and aggregating the style gradients of these crops with the global style gradients within one image. Furthermore, we propose a novel multi-objective optimization function to maximize visual discrepancy while enforcing semantic consistency among global, crop, and adversarial features. Applying the stabilized perturbations during training encourages convergence toward flatter and more transferable solutions, improving generalization to unseen domains. Extensive experiments are conducted on multiple CD-FSL benchmarks, demonstrating consistent improvements over state-of-the-art methods.

Problem

• Cross-Domain Few-Shot Learning
• style perturbation
• gradient instability
• domain shift
• sharp minima

Innovation

• Cross-Domain Few-Shot Learning
• Adversarial Style Perturbation
• Global-Crop Gradient Reorientation
• Semantic Consistency
• Flat Minima Optimization