https://github.com/mynameisvinn/federated-learning

tf implementation of federated learning
https://github.com/mynameisvinn/federated-learning

federated-learning machine-learning tensorflow

Last synced: about 2 months ago
JSON representation

tf implementation of federated learning

• Host: GitHub
• URL: https://github.com/mynameisvinn/federated-learning
• Owner: mynameisvinn
• Created: 2018-05-28T14:57:24.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2019-05-15T05:52:19.000Z (over 5 years ago)
• Last Synced: 2023-03-02T18:47:11.968Z (almost 2 years ago)
• Topics: federated-learning, machine-learning, tensorflow
• Language: Jupyter Notebook
• Size: 13.7 KB
• Stars: 36
• Watchers: 3
• Forks: 18
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# federated learning

a tensorflow implementation of ["federated learning: strategies for improving communication efficiency"](https://ai.google/research/pubs/pub45648).

the goal is to learn over distributed devices (eg smartphones), where each device holds data that may be (a) non iid, (b) imbalanced, and (c) sparse.

## stochastic variance reduced gradient (svrg)

[svrg](https://papers.nips.cc/paper/4937-accelerating-stochastic-gradient-descent-using-predictive-variance-reduction.pdf) is core to federated learning. when compared to vanilla sgd, it allows for faster convergence by reducing variance, introduced through small, noisy minibatches.