Summary of Scalable Federated Unlearning Via Isolated and Coded Sharding, by Yijing Lin et al.
Scalable Federated Unlearning via Isolated and Coded Sharding
by Yijing Lin, Zhipeng Gao, Hongyang Du, Dusit Niyato, Gui Gui, Shuguang Cui, Jinke Ren
First submitted to arxiv on: 29 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
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 scalable federated unlearning framework to address the limitations of current methods. Federated unlearning aims to erase client-level data effects without impacting collaborative learning model performance. However, existing approaches introduce significant storage overheads and computational resource consumption, hindering practical implementation. The authors develop an isolated sharding and coded computing-based framework to reduce affected clients and central server storage needs. They also provide theoretical analysis of time efficiency and storage effectiveness for the proposed method. Experimental results on classification and generation tasks demonstrate improved performance compared to three state-of-the-art frameworks in terms of accuracy, retraining time, storage overhead, and F1 scores for resisting membership inference attacks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps make artificial intelligence work better together by sharing knowledge without sharing personal data. Right now, when we want to remove old information from a model, it can be slow and use too much space. The researchers found a way to speed up the process and reduce storage needs. They divided clients into smaller groups and used special codes to compress information. This makes it easier to share knowledge without sharing personal data. The results show that this new method is better than others at keeping personal data safe while still getting good performance. |
