Summary of Data Pruning Can Do More: a Comprehensive Data Pruning Approach For Object Re-identification, by Zi Yang et al.
Data Pruning Can Do More: A Comprehensive Data Pruning Approach for Object Re-identification
by Zi Yang, Haojin Yang, Soumajit Majumder, Jorge Cardoso, Guillermo Gallego
First submitted to arxiv on: 13 Dec 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 proposed approach tackles data pruning in object re-identification (ReID) tasks, a problem long solved for image classification. By leveraging logit history during training, the method quantifies sample importance more accurately and corrects mislabeled samples while recognizing outliers. The framework is plug-and-play, architecture-agnostic, and reduces storage costs by eliminating/reducing 35%, 30%, and 5% of samples on VeRi, MSMT17, and Market1501 datasets, respectively, with negligible loss in accuracy (<0.1%). This approach is compared to existing methods, showing a 10-fold reduction in the cost of importance score estimation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A team of researchers has found a way to remove less important data from object recognition tasks. They used a special method that looks at how well each piece of data helps with recognizing objects. The new approach can get rid of up to 35% of unnecessary data, which makes training faster and uses fewer computer resources. This is an important step towards making object recognition more efficient. |
Keywords
» Artificial intelligence » Image classification » Pruning
