Summary of Cleft: Language-image Contrastive Learning with Efficient Large Language Model and Prompt Fine-tuning, by Yuexi Du et al.
CLEFT: Language-Image Contrastive Learning with Efficient Large Language Model and Prompt Fine-Tuning
by Yuexi Du, Brian Chang, Nicha C. Dvornek
First submitted to arxiv on: 30 Jul 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); 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 The paper introduces CLEFT, a novel contrastive learning method that combines pre-trained language and visual models for medical applications. It addresses the limitations of existing CLIP-like approaches by proposing an efficient strategy for prompt fine-tuning and leveraging clinical diagnostic data. The method achieves state-of-the-art performance on multiple chest X-ray and mammography datasets compared to various baselines, while reducing model size and trainable parameters. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary CLEFT is a new way to train language models that uses pictures and text together. This helps with medical tasks like analyzing X-rays. The old methods were slow and needed lots of computer power because they used big models and large amounts of data. CLEFT solves this problem by being more efficient and using prompts that are based on the information in the training samples, not just simple labels. This makes it better for real-world applications. |
Keywords
» Artificial intelligence » Fine tuning » Prompt
