Summary of Cross-training with Multi-view Knowledge Fusion For Heterogenous Federated Learning, by Zhuang Qi et al.
Cross-Training with Multi-View Knowledge Fusion for Heterogenous Federated Learning
by Zhuang Qi, Lei Meng, Weihao He, Ruohan Zhang, Yu Wang, Xin Qi, Xiangxu Meng
First submitted to arxiv on: 30 May 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 approach to federated learning, enhancing its generalization capability by incorporating multi-view information. Federated learning typically involves cross-training strategies to enable models to train on data from distinct sources. However, this process can lead to knowledge forgetting when adapting to new tasks or data sources. The proposed method, FedCT, consists of three modules: consistency-aware knowledge broadcasting, multi-view knowledge-guided representation learning, and mixup-based feature augmentation. These modules aim to optimize model assignment strategies, preserve local knowledge, and increase the diversity of feature spaces. Extensive experiments were conducted on four datasets, demonstrating that FedCT outperforms state-of-the-art methods by alleviating knowledge forgetting from both local and global views. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning helps machines learn together without sharing data. This makes it great for protecting privacy. However, this process can forget what was learned before. To solve this problem, researchers proposed a new way to do federated learning called FedCT. It has three parts: one that shares information between clients, another that keeps track of local knowledge, and another that adds more variety to the data. This helps the model learn better by keeping what it already knew and combining it with new information. |
Keywords
» Artificial intelligence » Federated learning » Generalization » Representation learning
