ViTAMINS: An Empirical Study of Training Self-Supervised Vision Transformers with Synthetic Hard Negatives

2026-09-01

AI Summary

本文提出ViTAMINS方法，通过在自监督视觉Transformer预训练中引入合成难负样本以提升表示质量，并在多个任务上展示了其优越性能。

Abstract

We introduce ViTAMINS, a method that integrates synthetic hard negatives into unsupervised vision transformer pretraining to improve representation quality. Our approach is thoroughly benchmarked on ImageNet and transfer learning, image retrieval, copy detection, and image, video segmentation tasks. Notably, our proposed negatives give rise to emergent properties, where learned representations contain explicit information about the semantic content of an image and serve as excellent classifiers (up to +11.3% over baselines). ViTAMINS achieves these benefits through simple modifications to existing contrastive frameworks and outperforms competing methods while being more resource efficient, e.g., our ViT-B surpasses V-JEPA with ViT-L. Our findings motivate reconsidering contrastive learning as a simpler yet powerful alternative to dominant generative and self-distillation approaches.

Problem

• self-supervised
• vision transformers
• synthetic hard negatives
• representation quality
• contrastive learning

Innovation

• synthetic hard negatives
• contrastive learning
• representation quality
• unsupervised vision transformer pretraining