Secure Collaborative Learning

Secure collaborative learning aims to enable multiple parties to jointly train machine learning models without compromising the privacy of their individual data or model weights. Current research focuses on developing efficient cryptographic techniques, such as hybrid homomorphic encryption and oblivious transfer, to enable secure computation and communication, as well as on mitigating adversarial attacks and data leakage through methods like data augmentation and unlearnable example generation. This field is crucial for facilitating data sharing in sensitive domains like healthcare and finance, while addressing growing concerns about data privacy and security in machine learning applications.