Summary of Labobf: a Label Protection Scheme For Vertical Federated Learning Through Label Obfuscation, by Ying He et al.
LabObf: A Label Protection Scheme for Vertical Federated Learning Through Label Obfuscation
by Ying He, Mingyang Niu, Jingyu Hua, Yunlong Mao, Xu Huang, Chen Li, Sheng Zhong
First submitted to arxiv on: 27 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: 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 Medium Difficulty summary: This paper addresses a crucial issue in vertical federated learning, where parties collaborate to improve model performance by sharing embedding vectors. The proposed Split Neural Network architecture allows for joint training without direct data exchange, but existing defenses can be bypassed using an embedding extension attack. To mitigate this threat, the authors introduce LabObf, a label obfuscation strategy that randomly maps integer-valued labels to multiple soft labels. By doing so, LabObf increases the difficulty for attackers to infer true labels while maintaining model accuracy. The paper conducts experiments on four datasets and shows that LabObf significantly reduces the attacker’s success rate compared to raw models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty summary: This research focuses on keeping data private in a special kind of teamwork called vertical federated learning. In this setup, different groups work together by sharing information that helps improve their model performance. The authors show that even though everyone is working together, some group members might still be able to guess the correct answers (labels) from the shared information. To fix this issue, they propose a new method called LabObf that makes it harder for attackers to guess the correct labels while keeping the overall model performance good. |
Keywords
» Artificial intelligence » Embedding » Federated learning » Neural network
