无需标签，用自监督学习混合模型参数提升鲁棒性。
Self-Soupervision: Cooking Model Soups without Labels
- 用自监督学习生成不同模型参数作为原料，混合成新模型。
- 在损坏数据上训练后，在正常数据上微调，鲁棒性提升3.5%~7%。
- 可混合不同自监督算法的参数，适合追求模型稳健性的研究者。
模型汤是一种奇特但高效的方法：以一个模型为基础（汤底），通过微调生成多个模型（食材），再将参数混合形成新模型（汤）。现有方法依赖有标签数据和监督学习，而本文提出自监督模型汤（Self-Souping），使其适用于无标签数据。该方法可在新数据源（如未标注任务数据或分布偏移数据）上生成多样化食材。实验显示，在损坏测试数据上进行自监督训练后，再在干净训练数据上微调，使ImageNet-C鲁棒性提升+3.5%，LAION-C提升+7%。此外，首次证明食材可采用不同自监督超参数甚至不同算法（如MAE、MoCoV3、MMCR、LeJEPA），混合后的模型精度优于任一单一算法。
原文摘要 · Abstract (English)
Model soups are strange and strangely effective combinations of parameters. They take a model (the stock), fine-tune it into multiple models (the ingredients), and then mix their parameters back into one model (the soup) to improve predictions. While all known soups require supervised learning, and optimize the same loss on labeled data, our recipes for Self-Soupervision generalize soups to self-supervised learning (SSL). Our Self-Souping lets us flavor ingredients on new data sources, e.g. from unlabeled data from a task for transfer or from a shift for robustness. We show that Self-Souping on corrupted test data, then fine-tuning back on uncorrupted train data, boosts robustness by +3.5% (ImageNet-C) and +7% (LAION-C). Self-Soupervision also unlocks countless SSL algorithms to cook the diverse ingredients needed for more robust soups. We show for the first time that ingredients can differ in their SSL hyperparameters -- and more surprisingly, in their SSL algorithms. We cook soups of MAE, MoCoV3, MMCR, and LeJEPA ingredients that are more accurate than any single SSL ingredient.
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。