Summary of Signsgd with Federated Voting, by Chanho Park et al.
SignSGD with Federated Voting
by Chanho Park, H. Vincent Poor, Namyoon Lee
First submitted to arxiv on: 25 Mar 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Distributed, Parallel, and Cluster Computing (cs.DC); Signal Processing (eess.SP)
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 paper proposes a novel distributed learning algorithm called signSGD-FV (federated voting) to accelerate model training on edge devices. The existing signSGD-MV algorithm reduces communication costs through one-bit quantization, but it fails to converge when workers have heterogeneous mini-batch sizes. To address this issue, the proposed signSGD-FV algorithm uses learnable weights to perform weighted majority voting, which are learned by the server based on edge devices’ computational capabilities. The paper provides a unified convergence rate analysis framework and demonstrates that signSGD-FV outperforms signSGD-MV, especially in heterogeneous mini-batch sizes. The proposed method has a theoretical convergence guarantee, making it suitable for practical applications. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research proposes a new way to make computer models learn faster on many devices at the same time. When devices share information with each other, it can take a long time because of how much data needs to be sent back and forth. The proposed method, called signSGD-FV, uses weights that are learned by the main server based on each device’s ability to process information. This helps devices with different processing powers work together better and makes the learning process faster. |
Keywords
* Artificial intelligence * Quantization
