Summary of Personalized Federated Learning Via Sequential Layer Expansion in Representation Learning, by Jaewon Jang and Bonjun Choi
Personalized Federated Learning via Sequential Layer Expansion in Representation Learning
by Jaewon Jang, Bonjun Choi
First submitted to arxiv on: 27 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 approach to address the issue of heterogeneity in federated learning by decoupling deep learning models into base and head components. The base component captures common features across all clients, while the head component captures unique features specific to individual clients. A new representation learning-based method is introduced, which divides the model into more densely divided parts with suitable scheduling methods. This approach benefits both data heterogeneity and class heterogeneity among clients. Two layer scheduling approaches, forward (Vanilla) and backward (Anti), are compared in the context of data and class heterogeneity. Experimental results show that the proposed algorithm achieves increased accuracy under challenging conditions while reducing computation costs, outperforming existing personalized federated learning algorithms. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making sure computer models learn from different people’s data without sharing too much information. They’re trying to solve a problem where the data is very different between people. The solution is to divide the model into two parts: one that is shared with everyone and another that is unique to each person. This helps the model work better when it sees different types of data or classes. Two ways to schedule this process are tested, and the results show that the new method works better than others in certain situations. |
Keywords
» Artificial intelligence » Deep learning » Federated learning » Representation learning
