通过分离数据流形内外扰动，提升迁移学习在域偏移下的鲁棒性。
Geometrically Regularized Transfer Learning with On-Manifold and Off-Manifold Perturbation
- 将对抗扰动分解为流形内与流形外两部分，分别捕捉语义变化与模型脆弱性。
- 在DomainNet等三个数据集上，比现有方法显著提升跨域泛化性能。
- 适合关注模型鲁棒性与领域自适应的科研人员和工程师。
由于源域与目标域数据流形的差异，域偏移下的迁移学习仍是根本挑战。本文提出MAADA（流形感知对抗数据增强）框架，将对抗扰动分解为流形内与流形外成分，同时捕捉语义变化与模型脆弱性。理论上证明，强制流形内一致性可降低假设复杂度并提升泛化能力，而流形外正则化则在低密度区域平滑决策边界。此外，引入几何感知对齐损失，最小化源域与目标域流形间的测地线差异。在DomainNet、VisDA和Office-Home数据集上的实验表明，MAADA在无监督和少样本设置下均持续优于现有对抗与自适应方法，展现出更强的结构鲁棒性与跨域泛化能力。
原文摘要 · Abstract (English)
Transfer learning under domain shift remains a fundamental challenge due to the divergence between source and target data manifolds. In this paper, we propose MAADA (Manifold-Aware Adversarial Data Augmentation), a novel framework that decomposes adversarial perturbations into on-manifold and off-manifold components to simultaneously capture semantic variation and model brittleness. We theoretically demonstrate that enforcing on-manifold consistency reduces hypothesis complexity and improves generalization, while off-manifold regularization smooths decision boundaries in low-density regions. Moreover, we introduce a geometry-aware alignment loss that minimizes geodesic discrepancy between source and target manifolds. Experiments on DomainNet, VisDA, and Office-Home show that MAADA consistently outperforms existing adversarial and adaptation methods in both unsupervised and few-shot settings, demonstrating superior structural robustness and cross-domain generalization.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。