Summary of Spear:exact Gradient Inversion Of Batches in Federated Learning, by Dimitar I. Dimitrov et al.
SPEAR:Exact Gradient Inversion of Batches in Federated Learning
by Dimitar I. Dimitrov, Maximilian Baader, Mark Niklas Müller, Martin Vechev
First submitted to arxiv on: 6 Mar 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Cryptography and Security (cs.CR); 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 research proposes SPEAR, a novel algorithm for reconstructing whole batches in federated learning when batch sizes are greater than one. The existing attacks only allow for approximate reconstruction of larger batches, but SPEAR leverages insights into the low-rank structure of gradients and ReLU-induced gradient sparsity to filter out incorrect samples, making exact reconstruction possible. The algorithm is implemented on a GPU for fully connected networks and successfully recovers high-dimensional ImageNet inputs in batches up to 25. Additionally, theoretical analysis suggests that larger batches can be reconstructed with high probability given exponential time. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary SPEAR is an algorithm that helps keep private data safe when people work together on machine learning projects without sharing their own information. Right now, there’s a problem where someone could reconstruct the original data from what they share. SPEAR solves this by finding the correct patterns in the shared information and filtering out the wrong parts. This makes it possible to get the original data back exactly, even with bigger groups of people working together. The researchers tested SPEAR on big images and showed that it can work well. |
Keywords
* Artificial intelligence * Federated learning * Machine learning * Probability * Relu
