Summary of Submfl: Compatiple Submodel Generation For Federated Learning in Device Heterogenous Environment, by Zeyneddin Oz et al.
subMFL: Compatiple subModel Generation for Federated Learning in Device Heterogenous Environment
by Zeyneddin Oz, Ceylan Soygul Oz, Abdollah Malekjafarian, Nima Afraz, Fatemeh Golpayegani
First submitted to arxiv on: 30 May 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 In this paper, researchers tackle a significant challenge in Federated Learning (FL), where devices with varying computing and storage capacities are unable to participate in the training process due to large model sizes. They propose a novel approach that compresses models to enable heterogeneous devices to join the training process, increasing participation rates by around 50%. This method uses dense models as initial global models for resource-constrained devices, which are then compressed to obtain submodels with varying levels of sparsity. The validation experiments show that these submodels maintain their accuracy despite reaching 50% global sparsity. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated Learning (FL) is a way for many devices to work together to make a shared model without sharing all their data. This helps keep the data private. But, big models like Deep Neural Networks need lots of computing power and energy. Some devices can’t handle that, so they’re left out. To fix this, scientists created a new way to compress models so even those small devices can join in. They do this by sharing a big model with all devices, then making it smaller step by step until it’s just right for each device. This helps more devices work together and keep their data safe. |
Keywords
* Artificial intelligence * Federated learning
