Private Machine Learning for Smart Energy Technologies
Protecting Privacy in Smart Energy with AI
Swinburne University of Technology and CSIRO’s Data61 are collaborating to address key challenges in the Artificial Intelligence sector.
Project Overview:
The project titled “Private Machine Learning for Smart Energy Technologies,” focuses on exploring smart energy technologies, while safeguarding people’s privacy and ensuring fairness. It seeks to promote ethical and responsible AI practices for smart energy analytics. This solution will make it both efficient and safe for Australians to utilise smart energy technologies.
Collaborative Approach:
Leveraging Swinburne University’s renowned expertise in Cyber Security for Machine Learning (Prof Jun Zhang and Prof Caslon Chua) and CSIRO’s industry leadership in Data Privacy for Artificial Intelligence, this project embodies a collaborative approach that bridges the gap between academia and industry.
Value for Industry:
Dr. Ming Ding (Industry Supervisor) highlighted the importance of the project, stating that it will develop new AI technologies for the smart energy sector, addressing ethical AI concerns such as users’ “right to be forgotten” and fair treatment of various population subgroups by AI. The project’s deliverables will fill a crucial gap in privacy protection and fairness enforcement in Australia’s efforts to reduce greenhouse gas emissions by 43% by 2030.
Value for Academia:
This project fosters technological innovations for advancing ethical AI such as federated machine unlearning algorithms that can retrain a near-optimal machine learning model after data deletion, and customised optimisation methods to balance the privacy preservation and fairness of machine learning models.
Student Perspective:
PhD Candidate Andrew Plapp, with a background in cyber security and machine learning, is excited to be taking part in this project. Andrew is excited at the opportunity to contribute new ideas to the field and hopes that the outcomes will create a solution that will protect Australian citizens’ data that is used by machine learning in the energy sector.
Project Outcomes/Impact
This project impact will allow for smart energy technologies in Australia using Machine Learning, to have the capabilities of removing users’ data at customers request.