Summary of Vision-language Models Are Strong Noisy Label Detectors, by Tong Wei and Hao-tian Li and Chun-shu Li and Jiang-xin Shi and Yu-feng Li and Min-ling Zhang
Vision-Language Models are Strong Noisy Label Detectors
by Tong Wei, Hao-Tian Li, Chun-Shu Li, Jiang-Xin Shi, Yu-Feng Li, Min-Ling Zhang
First submitted to arxiv on: 29 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
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 This paper proposes a Denoising Fine-Tuning (DeFT) framework for adapting vision-language models to downstream tasks while addressing the challenge of obtaining accurately labeled data. DeFT utilizes robust alignment of textual and visual features pre-trained on millions of auxiliary image-text pairs to sieve out noisy labels. The framework establishes a noisy label detector by learning positive and negative textual prompts for each class, promoting parameter-efficient fine-tuning of a pre-trained visual encoder. Experimental results validate the effectiveness of DeFT in both noisy label detection and image classification on seven synthetic and real-world datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps machines learn from images and text better. It’s hard to find good labels when training these models because labels can be wrong or missing. The researchers created a new way to fine-tune these models, called DeFT, that gets rid of bad labels. DeFT uses millions of image-text pairs to learn what good labels look like. Then, it finds the best words to use for each category and separates clean from noisy samples. This makes training more accurate and efficient. |
Keywords
» Artificial intelligence » Alignment » Encoder » Fine tuning » Image classification » Parameter efficient
