Summary of Privacy Attacks in Decentralized Learning, by Abdellah El Mrini et al.
Privacy Attacks in Decentralized Learning
by Abdellah El Mrini, Edwige Cyffers, Aurélien Bellet
First submitted to arxiv on: 15 Feb 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 This research proposes an attack on Decentralized Gradient Descent (D-GD), a collaborative learning method that allows users to learn without sharing their data. The attack enables users to reconstruct private data of others outside their immediate neighborhood by exploiting the gossip averaging protocol used in D-GD. The paper demonstrates the effectiveness of this attack on real graphs and datasets, showing that even a small number of attackers can compromise a large portion of the network. The results highlight the importance of developing robust defenses against these types of attacks to ensure secure collaborative learning. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Decentralized Gradient Descent is a way for people to learn together without sharing their data. But researchers found a way to hack into this system and steal information from others. They call it an attack, and it can affect many people if just a few hackers are involved. The study looked at different factors that make the attack work or not work, like how connected people are in the network, and how many hackers are trying to get in. |
Keywords
* Artificial intelligence * Gradient descent
