Summary of Vertical Federated Unlearning Via Backdoor Certification, by Mengde Han et al.
Vertical Federated Unlearning via Backdoor Certification
by Mengde Han, Tianqing Zhu, Lefeng Zhang, Huan Huo, Wanlei Zhou
First submitted to arxiv on: 16 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- 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 proposes a novel approach to Vertical Federated Learning (VFL) that enables entities to train models cooperatively while maintaining data privacy. The authors introduce an innovative mechanism to eliminate the influence of a specific client’s contribution within the VFL framework, which is crucial for addressing recent privacy regulations’ emphasis on individuals’ “right to be forgotten.” The method employs gradient ascent guided by a pre-defined constrained model to optimize model performance and avoid storing parameter updates or fully accessing the initial training data. The authors also introduce a backdoor mechanism to verify the effectiveness of the unlearning procedure. Empirical evidence shows that the results align closely with those achieved by retraining from scratch. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about a new way for companies or organizations to work together and share information without giving away their secrets. Imagine you’re working on a project with someone, but you want to make sure they don’t see all your notes. That’s basically what this paper is talking about – how to keep information private while still sharing it with others. The authors came up with a clever way to do this by using something called Vertical Federated Learning (VFL). It’s like a puzzle, where each person works on their part of the puzzle without seeing everyone else’s. They also figured out a way to “unlearn” some information that was already learned, which is important for keeping secrets safe. |
Keywords
» Artificial intelligence » Federated learning
