Summary of Rethinking the Role Of Pre-trained Networks in Source-free Domain Adaptation, by Wenyu Zhang et al.
Rethinking the Role of Pre-Trained Networks in Source-Free Domain Adaptation
by Wenyu Zhang, Li Shen, Chuan-Sheng Foo
First submitted to arxiv on: 15 Dec 2022
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 In this paper, researchers tackle the problem of source-free domain adaptation (SFDA), where a model trained on one dataset needs to adapt to another without labeled data. They propose a new approach that combines the strengths of pre-trained networks and traditional SFDA methods. The key idea is to integrate the pre-trained network into the target adaptation process, allowing it to provide an alternate view of features and classification decisions. This is achieved through a co-learning strategy that distills useful target domain information from the pre-trained network. The proposed approach outperforms existing methods on four benchmark datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The researchers developed a new method for adapting models trained on one dataset to another without labeled data. They used large pre-trained networks and a special way of training them together with the original model. This helped the model learn more about the new dataset and make better predictions. The results show that this approach is better than previous methods for doing this kind of adaptation. |
Keywords
* Artificial intelligence * Classification * Domain adaptation
