Summary of Learning Unlabeled Clients Divergence For Federated Semi-supervised Learning Via Anchor Model Aggregation, by Marawan Elbatel et al.
Learning Unlabeled Clients Divergence for Federated Semi-Supervised Learning via Anchor Model Aggregation
by Marawan Elbatel, Hualiang Wang, Jixiang Chen, Hao Wang, Xiaomeng Li
First submitted to arxiv on: 14 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
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 This paper proposes a novel federated semi-supervised learning (FedSemi) method called SemiAnAgg, which enables the aggregation of models from unlabeled clients in scenarios where there are clients with fully labeled, partially labeled, and even fully unlabeled data. The authors tackle the challenge of client drift due to heterogeneous class distributions and erroneous pseudo-labels by introducing an anchor model that learns the importance of each unlabeled client’s contribution. SemiAnAgg is demonstrated to achieve state-of-the-art results on four widely used FedSemi benchmarks, including a 9% increase in accuracy on CIFAR-100 and a 7.6% improvement in recall on the medical dataset ISIC-18. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper develops a new way for computers to learn from data when some of that data is not labeled. This is called federated semi-supervised learning (FedSemi). The problem is that some computers might have lots of labeled data, while others only have a little bit or no labels at all. To solve this issue, the authors created a new method called SemiAnAgg. It uses an anchor model to figure out which unlabeled clients are most helpful for learning. The results show that SemiAnAgg is much better than previous methods, and it can even improve accuracy by 9% on some datasets. |
Keywords
* Artificial intelligence * Recall * Semi supervised
