Summary of Federated Learning For Face Recognition Via Intra-subject Self-supervised Learning, by Hansol Kim et al.
Federated Learning for Face Recognition via Intra-subject Self-supervised Learning
by Hansol Kim, Hoyeol Choi, Youngjun Kwak
First submitted to arxiv on: 23 Jul 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); 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 A novel federated learning architecture called FedFS is proposed for personalized face recognition models without imposing subjects. The approach incorporates self-supervised learning and adaptive soft label construction to leverage aggregated features from local and global models. Additionally, intra-subject self-supervised learning strengthens robust representations using cosine similarity operations. Regularization loss prevents overfitting and ensures model stability. Experiments on DigiFace-1M and VGGFace datasets demonstrate superior performance compared to previous methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary FedFS is a new way to train face recognition models that works better than before. It uses self-learning and special label making to help local and global models work together. This helps make the model more robust and accurate for recognizing faces within a group or community. The approach also prevents overfitting and ensures the model stays stable. Tests on large datasets show that FedFS performs better than other methods. |
Keywords
» Artificial intelligence » Cosine similarity » Face recognition » Federated learning » Overfitting » Regularization » Self supervised
