Summary of Calibre: Towards Fair and Accurate Personalized Federated Learning with Self-supervised Learning, by Sijia Chen et al.
Calibre: Towards Fair and Accurate Personalized Federated Learning with Self-Supervised Learning
by Sijia Chen, Ningxin Su, Baochun Li
First submitted to arxiv on: 28 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
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 personalized federated learning framework called Calibre, designed to improve the performance of self-supervised learning (SSL) in heterogeneous data settings. In traditional SSL approaches, a global model is trained to extract transferable representations, allowing clients to train personalized models with limited data samples. However, when data is diverse across clients, the global model’s produced representations have fuzzy class boundaries, leading to low-accuracy personalized models. Calibre addresses this issue by calibrating SSL representations through client-specific prototype losses and an aggregation algorithm guided by these prototypes. The framework achieves state-of-the-art performance in terms of both mean accuracy and fairness across clients. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In simple terms, this paper is about improving how we do personalized learning on different devices using shared information. Right now, when the data is very different from one device to another, it’s hard to get good results. The researchers propose a new way called Calibre that can handle these differences and produce better results for each device. |
Keywords
» Artificial intelligence » Federated learning » Self supervised
