Summary of Pseudo-label Refinement For Improving Self-supervised Learning Systems, by Zia-ur-rehman et al.
Pseudo-label Refinement for Improving Self-Supervised Learning Systems
by Zia-ur-Rehman, Arif Mahmood, Wenxiong Kang
First submitted to arxiv on: 18 Oct 2024
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 The proposed pseudo-label refinement (SLR) algorithm addresses the issue of noise in clustering-based pseudo-labels by projecting cluster labels from previous epochs to the current epoch’s cluster-label space. This is done to better utilize information embedded in soft labels, which outperforms the common maximum value approach for hard label generation. The SLR algorithm achieves improved performance in person re-identification (Re-ID) tasks using unsupervised domain adaptation (UDA), with significant gains in mean Average Precision (mAP) across various scenarios. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a way to improve self-supervised learning systems by refining pseudo-labels. This helps the system learn better from clustering-based labels without human annotations. The new method, called SLR, uses the information from previous epochs and current epoch’s cluster labels to generate better hard labels. This leads to improved performance in tasks like recognizing people in different scenarios. |
Keywords
» Artificial intelligence » Clustering » Domain adaptation » Mean average precision » Self supervised » Unsupervised
