Summary of Feature-based Federated Transfer Learning: Communication Efficiency, Robustness and Privacy, by Feng Wang et al.
Feature-based Federated Transfer Learning: Communication Efficiency, Robustness and Privacy
by Feng Wang, M. Cenk Gursoy, Senem Velipasalar
First submitted to arxiv on: 15 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Multiagent Systems (cs.MA)
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 feature-based federated transfer learning approach reduces uplink payload by multiple orders of magnitude in federated learning and federated transfer learning. This method uploads extracted features and outputs instead of parameter updates, improving communication efficiency. Comparisons with existing schemes show the required payload is significantly reduced. The robustness against packet loss, data insufficiency, and quantization is analyzed, along with privacy considerations, including label and feature leakage. Mitigating approaches are investigated to ensure effective learning while preserving privacy. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper proposes a new way to make communication more efficient in machine learning. Instead of sending lots of information over the internet, we only send what’s necessary – features and outputs. This helps reduce the amount of data sent by many orders of magnitude compared to existing methods. We tested this approach on two tasks: image classification and natural language processing, and it worked well. |
Keywords
» Artificial intelligence » Federated learning » Image classification » Machine learning » Natural language processing » Quantization » Transfer learning
