/Collaborative-Incentives

Primary LanguageJupyter NotebookGNU General Public License v3.0GPL-3.0

Collaborative-Incentives

Experimental code mirroring what we used in One for One, or All for All:Equilibria and Optimality of Collaboration in Federated Learning. The original experiment was relatively computationally intensive, and so I've modified some of the iteration counts and the batch sizes to speed things up.

Convex Program and Optimization Code

The Finalnotebook.ipynb can be opened as a Jupyter notebook or in Google Colab to take a look at some of our various examples as well as programs to solve across certain different cases.