通过生成式随机化和跨变体自监督学习打破虚假相关性
Breaking Spurious Correlations via Generative Randomization and Cross-Variant Self-Supervised Learning
AI总结：
研究针对深度神经网络在分布转移时因利用虚假相关性而失败的问题，提出两阶段框架，先通过生成式干预分离前景对象并生成变体，再引入跨变体自监督学习，经微调后在多个基准测试中取得优异最差组性能。
摘要：
基于经验风险最小化（ERM）训练的深度神经网络在分布转移时常常失败，因为它们利用了对象标签与背景上下文之间的虚假相关性。近期生成式方法通过创建上下文改变的反事实图像解决此问题，但通常将这些样本用作标准数据增强，使模型仍可保留对背景敏感的表示。我们提出一个两阶段框架，利用生成式干预显式学习背景不变的视觉表示。首先，用零样本分割分离前景对象，并用结构保留扩散模型生成上下文转移变体，在改变周围环境时保留对象身份。接着引入跨变体自监督学习，不同背景下同一对象的变体在对比目标中形成正样本对，促使编码器对齐以对象为中心的表示并抑制背景特定线索。然后，使用ERM热身和具有逐层学习率的GroupDRO对预训练编码器进行微调。在分布转移基准测试上的实验展示了最佳的最差组性能，在水鸟数据集上达到92.5%，在MetaShift上达到81.7%，在NICO++上达到87.4%。
英文摘要：
Deep neural networks trained with Empirical Risk Minimization (ERM) often fail under distribution shifts because they exploit spurious correlations between object labels and background context. Recent generative approaches address this issue by creating counterfactual images with altered contexts, but typically use these samples as standard data augmentation, leaving the model free to retain background-sensitive representations. We propose a two-stage framework that uses generative intervention to explicitly learn background-invariant visual representations. First, we isolate the foreground object using zero-shot segmentation and generate context-shifted variants with a structure-preserving diffusion model, preserving object identity while varying the surrounding environment. We then introduce Cross-Variant Self-Supervised Learning, where variants of the same object under different backgrounds form positive pairs in a contrastive objective. This encourages the encoder to align object-centric representations while suppressing background-specific cues. Then, we fine-tune the pretrained encoder using an ERM warm-up followed by GroupDRO with layer-wise learning rates. Experiments on distribution-shift benchmarks demonstrate best worst-group performance, achieving 92.5% on Waterbirds, 81.7% on MetaShift, and 87.4% on NICO++. Code: https://github.com/surajyadav-research/GRSSL
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
