SecureAggregation

An implementation of Secure Aggregation algorithm based on "Practical Secure Aggregation for Privacy-Preserving Machine Learning (Bonawitz et. al)" in Python.

Dependencies: Flask, socketio and socketIO_client

pip install Flask

pip install socketio

pip install socketIO-client

Usage:

Client side:

Init:

c = secaggclient(host,port)

Give weights needed to be transmitted (originally set to zero)

c.setweights(ndnumpyarray,dimensionsofarray)

Set common base and mod

c.configure(commonbase, commonmod)

start client side:

c.start()

Server side:

init:

s = secaggserver(host,port,n,k)

where n is number of selected clients for the round and k is number of client responses required before aggregation process begins

start server side:

s.start()