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

• URL: http://arxiv.org/abs/2609.01041v1
• Date: Tue, 01 Sep 2026 10:39:25 GMT
• Title: ViTAMINS: An Empirical Study of Training Self-Supervised Vision Transformers with Synthetic Hard Negatives
• Abstract summary: We introduce ViTAMINS, a method that integrates synthetic hard negatives into unsupervised vision transformer pretraining to improve representation quality.<n>Our proposed negatives give rise to emergent properties, where learned representations contain explicit information about the semantic content of an image.<n>Our findings motivate reconsidering contrastive learning as a simpler yet powerful alternative to dominant generative and self-distillation approaches.
• Score: 5.571793666361683
• License: http://creativecommons.org/licenses/by/4.0/
• 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.

Related papers

• Unsupervised Training of Vision Transformers with Synthetic Negatives [5.571793666361683]
  We integrate synthetic hard negatives to improve vision transformer representation learning.<n>This simple yet effective technique notably improves the discriminative power of learned representations.
  arXiv  Detail & Related papers  (2025-09-02T07:14:21Z)
• Contrasting Intra-Modal and Ranking Cross-Modal Hard Negatives to Enhance Visio-Linguistic Compositional Understanding [6.798129852396113]
  We introduce a simple and effective method to improve compositional reasoning in Vision-Language Models (VLMs) Our method better leverages available datasets by refining and expanding the standard image-text contrastive learning framework. When integrated with CLIP, our technique yields notable improvement over state-of-the-art baselines.
  arXiv  Detail & Related papers  (2023-06-15T03:26:28Z)
• Self-Promoted Supervision for Few-Shot Transformer [178.52948452353834]
  Self-promoted sUpervisioN (SUN) is a few-shot learning framework for vision transformers (ViTs) SUN pretrains the ViT on the few-shot learning dataset and then uses it to generate individual location-specific supervision for guiding each patch token. Experiments show that SUN using ViTs significantly surpasses other few-shot learning frameworks with ViTs and is the first one that achieves higher performance than those CNN state-of-the-arts.
  arXiv  Detail & Related papers  (2022-03-14T12:53:27Z)
• Robust Contrastive Learning against Noisy Views [79.71880076439297]
  We propose a new contrastive loss function that is robust against noisy views. We show that our approach provides consistent improvements over the state-of-the-art image, video, and graph contrastive learning benchmarks.
  arXiv  Detail & Related papers  (2022-01-12T05:24:29Z)
• A Practical Contrastive Learning Framework for Single-Image Super-Resolution [51.422185656787285]
  We investigate contrastive learning-based single image super-resolution from two perspectives. We propose a practical contrastive learning framework for SISR, named PCL-SR. Compared with existing benchmark methods, we re-train them by our proposed PCL-SR framework and achieve superior performance.
  arXiv  Detail & Related papers  (2021-11-27T15:42:12Z)
• Self-Distilled Self-Supervised Representation Learning [35.60243157730165]
  State-of-the-art frameworks in self-supervised learning have recently shown that fully utilizing transformer-based models can lead to performance boost. In our work, we further exploit this by allowing the intermediate representations to learn from the final layers via the contrastive loss. Our method, Self-Distilled Self-Supervised Learning (SDSSL), outperforms competitive baselines (SimCLR, BYOL and MoCo v3) using ViT on various tasks and datasets.
  arXiv  Detail & Related papers  (2021-11-25T07:52:36Z)
• Efficient Self-supervised Vision Transformers for Representation Learning [86.57557009109411]
  We show that multi-stage architectures with sparse self-attentions can significantly reduce modeling complexity. We propose a new pre-training task of region matching which allows the model to capture fine-grained region dependencies. Our results show that combining the two techniques, EsViT achieves 81.3% top-1 on the ImageNet linear probe evaluation.
  arXiv  Detail & Related papers  (2021-06-17T19:57:33Z)
• Few-Shot Learning with Part Discovery and Augmentation from Unlabeled Images [79.34600869202373]
  We show that inductive bias can be learned from a flat collection of unlabeled images, and instantiated as transferable representations among seen and unseen classes. Specifically, we propose a novel part-based self-supervised representation learning scheme to learn transferable representations. Our method yields impressive results, outperforming the previous best unsupervised methods by 7.74% and 9.24%.
  arXiv  Detail & Related papers  (2021-05-25T12:22:11Z)
• Emerging Properties in Self-Supervised Vision Transformers [57.36837447500544]
  We show that self-supervised ViTs provide new properties to Vision Transformer (ViT) that stand out compared to convolutional networks (convnets) We implement our findings into a simple self-supervised method, called DINO, which we interpret as a form of self-distillation with no labels. We show the synergy between DINO and ViTs by achieving 80.1% top-1 on ImageNet in linear evaluation with ViT-Base.
  arXiv  Detail & Related papers  (2021-04-29T12:28:51Z)
• Contrastive Learning based Hybrid Networks for Long-Tailed Image Classification [31.647639786095993]
  We propose a novel hybrid network structure composed of a supervised contrastive loss to learn image representations and a cross-entropy loss to learn classifiers. Experiments on three long-tailed classification datasets demonstrate the advantage of the proposed contrastive learning based hybrid networks in long-tailed classification.
  arXiv  Detail & Related papers  (2021-03-26T05:22:36Z)
• A Broad Study on the Transferability of Visual Representations with Contrastive Learning [15.667240680328922]
  We study the transferability of learned representations of contrastive approaches for linear evaluation, full-network transfer, and few-shot recognition. The results show that the contrastive approaches learn representations that are easily transferable to a different downstream task. Our analysis reveals that the representations learned from the contrastive approaches contain more low/mid-level semantics than cross-entropy models.
  arXiv  Detail & Related papers  (2021-03-24T22:55:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.