Get Started
Home
Topics
Search
Library
Research questionHow can low-level vision generalist models adapt to previously unseen tasks from a single example?Low-level vision generalist models must transfer beyond the tasks they learned during training, but their ability to recognize and perform an unseen task from one example remains uncertain. Systematic assessment across new task types is also limited.
AI
Computer Vision
Evaluation & Benchmarks
Image & Video Processing
Machine Learning
Latest papersRecent research connected to this question, newest first.Hidden-Shot: Towards One-Shot Task Generalization for Low-Level Vision Generalist ModelsApplies to low-level vision generalist models and the one-shot adaptation setting studied in the source. Evidence covers seven datasets in the retraining setting and ten datasets when training from scratch, using the proposed C/U assessment framework; it does not establish performance beyond those evaluated datasets.research paper · Sep 2, 2026
Related questions
How can vision models trained on one labeled domain generalize to unseen domains with difficult, semantically valid variations?How can detect-to-track pipelines adapt to new visual domains without target-domain labels?How can vision-language models maintain visual recognition when modalities are missing and source training data is unavailable?Can small targeted grayscale patches force chosen semantics in infrared vision-language models across tasks?