Summary of Colafier: Collaborative Noisy Label Purifier with Local Intrinsic Dimensionality Guidance, by Dongyu Zhang et al.
CoLafier: Collaborative Noisy Label Purifier With Local Intrinsic Dimensionality Guidance
by Dongyu Zhang, Ruofan Hu, Elke Rundensteiner
First submitted to arxiv on: 10 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- 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 In this paper, researchers introduce CoLafier, a novel approach for learning with noisy labels. The method utilizes Local Intrinsic Dimensionality (LID) to predict label quality and assign weights to training instances. CoLafier consists of two subnets: LID-dis, which consumes features and labels to produce an enhanced internal representation, and LID-gen, a regular classifier that operates solely on features. During training, the model uses two augmented views per instance to feed both subnets. The approach demonstrates improved prediction accuracy under severe label noise, outperforming existing methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary CoLafier is a new way for computers to learn from noisy data. Noisy data means some of the information is incorrect or misleading. The researchers created CoLafier to help machines learn better by using something called Local Intrinsic Dimensionality (LID). LID helps figure out which parts of the data are trustworthy and which aren’t. This helps the machine make more accurate predictions. |
