mlsanityData to 14 Oct 2026

Learning Vision from Models Rivals Learning Vision from Data

arXiv 2312.17742 ↗First posted 28 Dec 2023cs.CVcited by 71

Not on the map. Why?

Abstract (arXiv)

We introduce SynCLR, a novel approach for learning visual representations exclusively from synthetic images and synthetic captions, without any real data. We synthesize a large dataset of image captions using LLMs, then use an off-the-shelf text-to-image model to generate multiple images corresponding to each synthetic caption. We perform visual representation learning on these synthetic images via contrastive learning, treating images sharing the same caption as positive pairs. The resulting representations transfer well to many downstream tasks, competing favorably with other general-purpose visual representation learners such as CLIP and DINO v2 in image classification tasks. Furthermore, in dense prediction tasks such as semantic segmentation, SynCLR outperforms previous self-supervised methods by a significant margin, e.g., improving over MAE and iBOT by 6.2 and 4.3 mIoU on ADE20k for ViT-B/16.

Key facts. Cited by 71 papers in the map (data through 14 Oct 2026). Builds on 89 of its 165 references that are in the map. First posted on arXiv 28 Dec 2023.

Builds on

89 of 165 references.

References · 89

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  3. 2020Language Models are Few-Shot Learnerscited by 32,419
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  5. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  6. 2023GPT-4 Technical Reportcited by 18,396
  7. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,364
  8. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  9. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  10. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  11. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  12. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,513
38 more
  1. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  2. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  3. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,355
  4. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,183
  5. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,750
  6. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,152
  7. 2021Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervisioncited by 3,864
  8. 2019CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Featurescited by 2,884
  9. 2020Unsupervised Learning of Visual Features by Contrasting Cluster Assignmentscited by 2,857
  10. 2020Supervised Contrastive Learningcited by 2,829
  11. 2015Learning Deep Features for Discriminative Localizationcited by 2,436
  12. 2021BEiT: BERT Pre-Training of Image Transformerscited by 2,375
  13. 2013Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distancescited by 2,316
  14. 2019RandAugment: Practical automated data augmentation with a reduced search spacecited by 2,269
  15. 2013Rich feature hierarchies for accurate object detection and semantic segmentationcited by 2,162
  16. 2013Describing Textures in the Wildcited by 2,125
  17. 2018Unsupervised Feature Learning via Non-Parametric Instance-level Discriminationcited by 1,845
  18. 2020GLU Variants Improve Transformercited by 1,804
  19. 2013Fine-Grained Visual Classification of Aircraftcited by 1,726
  20. 2021An Empirical Study of Training Self-Supervised Vision Transformerscited by 1,616
  21. 2018Unsupervised Representation Learning by Predicting Image Rotationscited by 1,591
  22. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,496
  23. 2020Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hyperspherecited by 1,405
  24. 2016Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzlescited by 1,382
  25. 2019Contrastive Multiview Codingcited by 1,301
  26. 2018Unified Perceptual Parsing for Scene Understandingcited by 1,271
  27. 2021Conceptual 12M: Pushing Web-Scale Image-Text Pre-Training To Recognize Long-Tail Visual Conceptscited by 1,216
  28. 2022Reproducible scaling laws for contrastive language-image learningcited by 1,173
  29. 2021SimMIM: A Simple Framework for Masked Image Modelingcited by 1,153
  30. 2016Colorful Image Colorizationcited by 1,114
  31. 2016Semantic Understanding of Scenes through the ADE20K Datasetcited by 1,069
  32. 2021Scaling Vision Transformerscited by 999
  33. 2016Deep Networks with Stochastic Depthcited by 979
  34. 2021iBOT: Image BERT Pre-Training with Online Tokenizercited by 903
  35. 2015A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimationcited by 881
  36. 2020What Makes for Good Views for Contrastive Learning?cited by 846
  37. 2017Remote Sensing Image Scene Classification: Benchmark and State of the Artcited by 797
  38. 2021Going deeper with Image Transformerscited by 764

50 most cited of 89 All 89 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 36; 2025: 24; 2026: 11.

71 papers.

Sort:

By area:

  1. 2024Depth Anything V2cited by 1,422
  2. 2024An Introduction to Vision-Language Modelingcited by 178
  3. 2024CinePile: A Long Video Question Answering Dataset and Benchmarkcited by 114
  4. 2024Manipulate-Anything: Automating Real-World Robots using Vision-Language Modelscited by 83
  5. 2025Vision-Language Models Do Not Understand Negationcited by 72
  6. 2024SynthCLIP: Are We Ready for a Fully Synthetic CLIP Training?cited by 64
  7. 2024Self-Correcting Self-Consuming Loops for Generative Model Trainingcited by 37
  8. 2024TIPS: Text-Image Pretraining with Spatial awarenesscited by 37
  9. 2026Orient Anything V2: Unifying Orientation and Rotation Understandingcited by 23
  10. 2024The Unmet Promise of Synthetic Training Images: Using Retrieved Real Images Performs Bettercited by 19
  11. 2024Minority-Focused Text-to-Image Generation via Prompt Optimizationcited by 18
  12. 2024When Does Perceptual Alignment Benefit Vision Representations?cited by 18
38 more
  1. 2024Objective drives the consistency of representational similarity across datasetscited by 18
  2. 2025Improving the Scaling Laws of Synthetic Data with Deliberate Practicecited by 17
  3. 2024Understanding Bias in Large-Scale Visual Datasetscited by 15
  4. 2025A Large Scale Analysis of Gender Biases in Text-to-Image Generative Modelscited by 14
  5. 2024Detecting Human Artifacts from Text-to-Image Modelscited by 12
  6. 2024Efficient Lifelong Model Evaluation in an Era of Rapid Progresscited by 10
  7. 2024Accessing Vision Foundation Models via ImageNet-1Kcited by 10
  8. 2024Can OOD Object Detectors Learn from Foundation Models?cited by 10
  9. 2025How far can we go with ImageNet for Text-to-Image generation?cited by 10
  10. 2024Personalized Representation from Personalized Generationcited by 9
  11. 2024Can Medical Vision-Language Pre-training Succeed with Purely Synthetic Data?cited by 7
  12. 2025Human-Aligned Image Models Improve Visual Decoding from the Braincited by 7
  13. 2025Logic Unseen: Revealing the Logical Blindspots of Vision-Language Modelscited by 7
  14. 2025Bongard-RWR+: Real-World Representations of Fine-Grained Concepts in Bongard Problemscited by 7
  15. 2024FinLLMs: A Framework for Financial Reasoning Dataset Generation with Large Language Modelscited by 6
  16. 2026When Pretty Isn't Useful: Investigating Why Modern Text-to-Image Models Fail as Reliable Training Data Generatorscited by 6
  17. 2024Are Images Indistinguishable to Humans Also Indistinguishable to Classifiers?cited by 5
  18. 2025Towards Generating Realistic 3D Semantic Training Data for Autonomous Drivingcited by 5
  19. 2024Learning from Synthetic Data for Visual Groundingcited by 4
  20. 2024DODA: Adapting Object Detectors to Dynamic Agricultural Environments in Real-Time with Diffusioncited by 4
  21. 2024Zero-Shot Distillation for Image Encoders: How to Make Effective Use of Synthetic Datacited by 4
  22. 2024Robust Disaster Assessment from Aerial Imagery Using Text-to-Image Synthetic Datacited by 4
  23. 2024Auto Cherry-Picker: Learning from High-quality Generative Data Driven by Languagecited by 4
  24. 2025ClaraVid: A Holistic Scene Reconstruction Benchmark From Aerial Perspective With Delentropy-Based Complexity Profilingcited by 4
  25. 2025Aligning Generative Denoising with Discriminative Objectives Unleashes Diffusion for Visual Perceptioncited by 4
  26. 2024Weak-to-Strong Compositional Learning from Generative Models for Language-based Object Detectioncited by 3
  27. 2026Learning Brain Representation with Hierarchical Visual Embeddingscited by 3
  28. 2024ComFace: Facial Representation Learning with Synthetic Data for Comparing Facescited by 2
  29. 2024Mask-guided cross-image attention for zero-shot in-silico histopathologic image generation with a diffusion modelcited by 2
  30. 2024Vision CNNs trained to estimate spatial latents learned similar ventral-stream-aligned representationscited by 2
  31. 2025LoRA-Loop: Closing the Synthetic Replay Cycle for Continual VLM Learningcited by 2
  32. 2026Dance Across Shifts: Forward-Facilitation Continual Test-Time Adaptation through Dynamic Style Bridgingcited by 2
  33. 2026On the Difficulty of Learning a Meta-network for Training Data Selectioncited by 2
  34. 2024Contrastive Learning from Synthetic Audio Doppelg"angerscited by 1
  35. 2024Precision at Scale: Domain-Specific Datasets On-Demandcited by 1
  36. 2024Paint Outside the Box: Synthesizing and Selecting Training Data for Visual Groundingcited by 1
  37. 2025Search is All You Need for Few-shot Anomaly Detectioncited by 1
  38. 2025LoFT: LoRA-fused Training Dataset Generation with Few-shot Guidancecited by 1

50 most cited of 71 All 71 (JSON)