Ensuring reliable deployment of deep neural network models. This project aims to develop novel techniques for the reliable deployment of deep neural network models in machine learning as a service. It

Description

Ensuring reliable deployment of deep neural network models. This project aims to develop novel techniques for the reliable deployment of deep neural network models in machine learning as a service. It addresses the important and urgent problem of how to ensure secure and flexible model access and robust model infringement tracing. The project will result in a set of innovative methods involving deep neural network model access control, model piracy detection and model ownership identification. This should provide significant benefits by empowering Australia's leading position in artificial intelligence security research, enhancing the competitiveness of the local artificial intelligence industry, and preventing cyber crimes and huge financial and job losses caused by model infringements.. Scheme: Linkage Projects. Field: 4604 - Cybersecurity and Privacy. Lead: Prof Yong Xiang