Summary of Unmasking Efficiency: Learning Salient Sparse Models in Non-iid Federated Learning, by Riyasat Ohib et al.
Unmasking Efficiency: Learning Salient Sparse Models in Non-IID Federated Learning
by Riyasat Ohib, Bishal Thapaliya, Gintare Karolina Dziugaite, Jingyu Liu, Vince Calhoun, Sergey Plis
First submitted to arxiv on: 15 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); 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 The proposed Salient Sparse Federated Learning (SSFL) approach streamlines sparse federated learning with efficient communication. By identifying a sparse subnetwork prior to training, leveraging parameter saliency scores computed on local client data, and aggregating them globally, only the sparse model weights are communicated between clients and the server. This method shows marked improvements in sparsity-accuracy trade-offs when validated using standard non-IID benchmarks. The effectiveness of SSFL is also demonstrated in a real-world federated learning framework, resulting in improved communication time. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning helps computers learn together without sharing all their data. This paper proposes a new way to make this process more efficient by only sending parts of the information that’s most important. It uses special scores to figure out which parts are most important and sends just those, making it faster and better than before. |
Keywords
» Artificial intelligence » Federated learning
