Summary of Ot-vp: Optimal Transport-guided Visual Prompting For Test-time Adaptation, by Yunbei Zhang et al.
OT-VP: Optimal Transport-guided Visual Prompting for Test-Time Adaptation
by Yunbei Zhang, Akshay Mehra, Jihun Hamm
First submitted to arxiv on: 12 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel approach, Optimal Transport-guided Test-Time Visual Prompting (OT-VP), addresses the limitations of Vision Transformers (ViTs) in unseen domains. Unlike previous methods, OT-VP learns a universal visual prompt at test time without altering pre-trained model parameters or accessing training data. This prompts learning, optimized using Optimal Transport, enables ViT models to align target and source domains effectively. The method demonstrates state-of-the-art performance on three stylistic datasets (PACS, VLCS, OfficeHome) and one corrupted dataset (ImageNet-C), using only four learned prompt tokens. OT-VP also operates efficiently in terms of memory and computation, making it suitable for online settings. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new way to help computers understand pictures better is discovered. This method, called Optimal Transport-guided Test-Time Visual Prompting, makes sure that computers can understand new types of pictures even if they haven’t seen them before. It does this by learning a special set of instructions, or “prompts,” that the computer can use to look at new pictures in the right way. This method works well and is fast, making it useful for computers to learn about new things. |
Keywords
» Artificial intelligence » Prompt » Prompting » Vit
