Summary of Personalized Federated Domain-incremental Learning Based on Adaptive Knowledge Matching, by Yichen Li et al.
Personalized Federated Domain-Incremental Learning based on Adaptive Knowledge Matching
by Yichen Li, Wenchao Xu, Haozhao Wang, Ruixuan Li, Yining Qi, Jingcai Guo
First submitted to arxiv on: 6 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: 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 approach, adaptive knowledge matching-based personalized Federated Domain-Incremental Learning (pFedDIL), which allows each client to utilize the most suitable incremental task learning strategy based on the correlation with their previous tasks. In pFedDIL, clients first calculate local correlations between new and previous tasks, then choose an initial model or a previous model with similar knowledge to train the new task. The approach also condenses model parameters by sharing partial parameters between target classification models and auxiliary classifiers. Experimental results show that pFedDIL outperforms state-of-the-art methods by up to 14.35% in terms of average accuracy across all tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary FDIL is a type of learning where each client learns new tasks as their domain shifts from one another. The paper proposes a new way for clients to learn these tasks, called pFedDIL. This method helps clients choose the best way to train new tasks based on how similar they are to previous tasks. It also helps clients share knowledge between old and new tasks. The results show that this approach works better than other methods. |
Keywords
* Artificial intelligence * Classification
