Matrix Encryption

Matrix encryption is a cryptographic technique increasingly used to enhance the privacy of data in collaborative machine learning settings, particularly federated learning. Current research focuses on applying matrix encryption to various models, such as logistic regression and stochastic gradient descent, to protect sensitive information during training while maintaining model accuracy and efficiency. This approach addresses concerns about data leakage in distributed systems, offering a promising solution for secure collaborative data analysis in applications ranging from Internet of Things data processing to sensitive scientific collaborations. The ultimate goal is to balance strong privacy guarantees with minimal impact on model performance and computational overhead.

Papers

April 5, 2022

Privacy-Preserving Federated Learning via System Immersion and Random Matrix Encryption
Haleh Hayati, Carlos Murguia, Nathan van de Wouw
Privacy Preserving Privacy Preserving Federated Learning Immersive Environment Matrix Encryption

February 5, 2022

January 26, 2022

Speckle-based optical cryptosystem and its application for human face recognition via deep learning
Qi Zhao, Huanhao Li, Zhipeng Yu, Chi Man Woo, Tianting Zhong, Shengfu Cheng, Yuanjin Zheng, Honglin Liu, Jie Tian, Puxiang Lai
Deep Learning Application Proficiency Face Recognition Face Image Speckle Noise Matrix Encryption

Matrix Encryption

Papers

Privacy-Preserving Federated Learning via System Immersion and Random Matrix Encryption

Efficient Privacy Preserving Logistic Regression for Horizontally Distributed Data

Linear Model Against Malicious Adversaries with Local Differential Privacy

Speckle-based optical cryptosystem and its application for human face recognition via deep learning