Professor Jie Lu contributes to United Nations working group

Held on 8 and 9 September, the Working Group is developing recommendations towards equitable and interoperable data governance arrangements for consideration by the United Nations General Assembly.

Professor Lu participated as an observer and delivered an intervention supporting the Working Group’s proposals on interoperability between national, regional and international data systems. Drawing on AAII’s expertise in artificial intelligence and data science, Prof Lu highlighted practical ways in which AI technologies can translate interoperability principles into implementable solutions.

A three-pronged approach 

The AAII intervention proposed a three-pronged approach combining Open Science, Transfer Learning and Federated Learning. It advocated making non-sensitive research data and reusable research assets publicly accessible through open platforms such as GitHub and Kaggle.

The intervention also proposed applying transfer learning to reuse knowledge from mature, data-rich environments in data-scarce populations and jurisdictions. For example, adapting models trained on large international genomic datasets to smaller Australian cohorts, such as in women’s health applications. Federated learning can enable institutions to train or analyse models collaboratively while sensitive data remains within secure local environments.

Providing a practical pathway towards data interoperability that is open where possible, privacy-preserving where necessary, and designed for reuse across heterogeneous data environments 

Together, these approaches provide a practical pathway towards data interoperability that is open where possible, privacy-preserving where necessary, and designed for reuse across heterogeneous data environments. 

Professor Lu’s intervention demonstrated how AAII’s capabilities in responsible AI can contribute technical solutions to the international discussion on sustainable and trustworthy data governance.