mlsanityData to 7 Oct 2026

Test-Time Visual In-Context Tuning

arXiv 2503.21777 ↗First posted 27 Mar 2025cs.CV cs.LGcited by 7

Not on the map. Why?

Abstract (arXiv)

Visual in-context learning (VICL), as a new paradigm in computer vision, allows the model to rapidly adapt to various tasks with only a handful of prompts and examples. While effective, the existing VICL paradigm exhibits poor generalizability under distribution shifts. In this work, we propose test-time Visual In-Context Tuning (VICT), a method that can adapt VICL models on the fly with a single test sample. Specifically, we flip the role between the task prompts and the test sample and use a cycle consistency loss to reconstruct the original task prompt output. Our key insight is that a model should be aware of a new test distribution if it can successfully recover the original task prompts. Extensive experiments on six representative vision tasks ranging from high-level visual understanding to low-level image processing, with 15 common corruptions, demonstrate that our VICT can improve the generalizability of VICL to unseen new domains. In addition, we show the potential of applying VICT for unseen tasks at test time. Code: https://github.com/Jiahao000/VICT.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 52 of its 66 references that are in the map. First posted on arXiv 27 Mar 2025.

Builds on

52 of 66 references.

References · 52

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Language Models are Few-Shot Learnerscited by 32,199
  3. 2023GPT-4 Technical Reportcited by 18,132
  4. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  5. 2024The Llama 3 Herd of Modelscited by 14,192
  6. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  7. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
  8. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  9. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,339
  10. 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 4,337
  11. 2024Gemini 1.5: Unlocking multimodal understanding across millions of tokens of contextcited by 3,175
  12. 2023Sigmoid Loss for Language Image Pre-Trainingcited by 3,048
38 more
  1. 2022Scaling Instruction-Finetuned Language Modelscited by 2,997
  2. 2015Fast R-CNNcited by 2,930
  3. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,857
  4. 2018Unsupervised Representation Learning by Predicting Image Rotationscited by 1,586
  5. 2019Theoretically Principled Trade-off between Robustness and Accuracycited by 1,510
  6. 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,401
  7. 2016Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzlescited by 1,379
  8. 2015Unsupervised Visual Representation Learning by Context Predictioncited by 1,062
  9. 2018Robustness May Be at Odds with Accuracycited by 873
  10. 2016Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurringcited by 722
  11. 2018Deep Retinex Decomposition for Low-Light Enhancementcited by 720
  12. 2018Taskonomy: Disentangling Task Transfer Learningcited by 631
  13. 2022Test-Time Prompt Tuning for Zero-Shot Generalization in Vision-Language Modelscited by 379
  14. 2019Benchmarking Robustness in Object Detection: Autonomous Driving when Winter is Comingcited by 332
  15. 2022Images Speak in Images: A Generalist Painter for In-Context Visual Learningcited by 317
  16. 2020Learning Enriched Features for Real Image Restoration and Enhancementcited by 311
  17. 2022Contrastive Test-Time Adaptationcited by 258
  18. 2022Visual Prompting via Image Inpaintingcited by 248
  19. 2023Sequential Modeling Enables Scalable Learning for Large Vision Modelscited by 212
  20. 2022Test-Time Training with Masked Autoencoderscited by 208
  21. 2016Improving the Robustness of Deep Neural Networks via Stability Trainingcited by 156
  22. 2023Diverse Data Augmentation with Diffusions for Effective Test-time Prompt Tuningcited by 150
  23. 2024Efficient Test-Time Adaptation of Vision-Language Modelscited by 138
  24. 2020Self-Supervised Policy Adaptation during Deploymentcited by 121
  25. 2020Online Deep Clustering for Unsupervised Representation Learningcited by 99
  26. 2022Language Models are General-Purpose Interfacescited by 96
  27. 2018Real-time self-adaptive deep stereocited by 80
  28. 2022Masked Frequency Modeling for Self-Supervised Visual Pre-Trainingcited by 70
  29. 2023SILC: Improving Vision Language Pretraining with Self-Distillationcited by 66
  30. 2021MT3: Meta Test-Time Training for Self-Supervised Test-Time Adaptioncited by 65
  31. 2024On the test-time zero-shot generalization of vision-language models: Do we really need prompt learning?cited by 65
  32. 2016Examining the Impact of Blur on Recognition by Convolutional Networkscited by 58
  33. 2020Delving into Inter-Image Invariance for Unsupervised Visual Representationscited by 37
  34. 2020Continual Adaptation for Deep Stereocited by 23
  35. 2021Self-Supervised Test-Time Learning for Reading Comprehensioncited by 22
  36. 2021Learning to Track Instances without Video Annotationscited by 18
  37. 2023Convolutional Visual Prompt for Robust Visual Perceptioncited by 17
  38. 2017Quality Resilient Deep Neural Networkscited by 12

50 most cited of 52 All 52 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 3; 2026: 4.

7 papers.

  1. 2025Video models are zero-shot learners and reasonerscited by 208
  2. 2025Prompt-based Adaptation in Large-scale Vision Models: A Surveycited by 21
  3. 2025CNS-Bench: Benchmarking Image Classifier Robustness Under Continuous Nuisance Shiftscited by 2
  4. 2026ClipTTT: CLIP-Guided Test-Time Training Helps LVLMs See Bettercited by 2
  5. 2026Beyond Single Prompts: Synergistic Fusion and Arrangement for VICLcited by 0
  6. 2026Quo Vadis, Visual In-Context Learning? A Unified Benchmark Across Domains and Taskscited by 0
  7. 2026When Should the Count Change? Learning State Maintenance for Causal Video Countingcited by 0