Summary of Personalized Federated Learning Via Backbone Self-distillation, by Pengju Wang and Bochao Liu and Dan Zeng and Chenggang Yan and Shiming Ge
Personalized Federated Learning via Backbone Self-Distillation
by Pengju Wang, Bochao Liu, Dan Zeng, Chenggang Yan, Shiming Ge
First submitted to arxiv on: 24 Sep 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); 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 presents a novel method to facilitate personalized federated learning in practical scenarios where clients have heterogeneous data. The authors propose a backbone self-distillation approach, which trains local models for each client using only the backbone weights sent from the server. To address the lack of personalization in the client’s local backbone, they use the global backbone as a teacher to update the local backbone through a process called backbone self-distillation. This process involves learning two components: a shared backbone for common representation and a private head for local personalization. The authors demonstrate the effectiveness of their approach through extensive experiments and comparisons with 12 state-of-the-art methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In this paper, scientists developed a new way to make personalized models work better together when they’re trained on different data. They call it backbone self-distillation. It works by having each client’s model learn from the global model, which is like a teacher. This helps the local model become more personalized and good at learning things that are unique to its own data. The scientists tested their idea with many other approaches and showed that it works well. |
Keywords
» Artificial intelligence » Distillation » Federated learning
