This repository includes code and results reproduced from the paper "Distributed Differentially-Private Canonical Correlation Analysis" by Hafiz Imtiaz and Anand D. Sarwate. This paper assumes an 'honest but curious' decentralized setting.
Please refer to the original paper: https://ieeexplore.ieee.org/document/8683252
Please send bug reports and requests for tutorials to the following contact.
Contact: [email protected]
The following is a list of programs:
- local.py - Simulating a local site for collaboration
- remote.py - Simulating a remote site for collaboration
- generate_data.py, generate_data_mnist.py, generate_data_xrmb.py - Generates synthetic data for testing of algorithm
- compute_utilities.py - Computes performance metrics as described in the paper