Summary of Contribution-based Low-rank Adaptation with Pre-training Model For Real Image Restoration, by Donwon Park et al.
Contribution-based Low-Rank Adaptation with Pre-training Model for Real Image Restoration
by Donwon Park, Hayeon Kim, Se Young Chun
First submitted to arxiv on: 2 Aug 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
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 proposes a novel efficient parameter tuning approach, Contribution-Based Low-Rank Adaptation (CoLoRA), for multiple image restorations. CoLoRA fine-tunes only a small amount of parameters by leveraging LoRA to adaptively determine layer-by-layer capacity for each new task. The approach is compared to full tuning and achieves comparable performance while being more efficient. Additionally, the paper introduces a pre-training method called Random Order Degradations (PROD) that extends the capability of pre-trained models with improved performance and robustness. The authors demonstrate the effectiveness of CoLoRA with PROD on various image restoration tasks across diverse degradation types using both synthetic and real-world datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Imagine you have a superpower that helps machines learn new skills quickly and efficiently. This paper is all about developing that superpower for machines to get better at restoring images, like fixing blurry pictures or removing noise from old photos. The researchers came up with two cool ideas: one called CoLoRA (say “koh-loh-ruh”) that helps machines adjust their learning process to fit new tasks, and another called PROD (short for “random order degradations”) that teaches machines how to learn from imperfect training data. By combining these two ideas, the researchers showed that machines can get really good at restoring images without needing too much training or computer power. |
Keywords
» Artificial intelligence » Lora » Low rank adaptation
