Summary of One-shot Federated Learning Via Synthetic Distiller-distillate Communication, by Junyuan Zhang et al.
One-shot Federated Learning via Synthetic Distiller-Distillate Communication
by Junyuan Zhang, Songhua Liu, Xinchao Wang
First submitted to arxiv on: 6 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: 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 In this paper, researchers propose a novel one-shot federated learning (FL) framework called FedSD2C to address the challenges of data heterogeneity and information loss in FL. One-shot FL is an efficient approach that can train machine learning models with a single round of communication, but it often compromises model performance. The proposed framework introduces a distiller to synthesize informative distillates directly from local data, reducing information loss and tackling data heterogeneity by sharing synthetic distillates instead of inconsistent local models. Experimental results show that FedSD2C outperforms other one-shot FL methods on complex and real-world datasets, achieving up to 2.6 times the performance of the best baseline. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning (FL) is a way for many devices to work together and learn from each other’s data without sharing their individual information. One-shot FL is faster than traditional FL because it only needs one round of communication, but it can be less accurate. The researchers in this paper want to make one-shot FL better by reducing the loss of important information when sharing knowledge between devices. They propose a new method called FedSD2C that can do this and show that it works well on real-world data. |
Keywords
* Artificial intelligence * Federated learning * Machine learning * One shot
