ÌÇÐÄvlog¹ÙÍø

Artificial intelligence (AI) is one of the defining technologies of the 21st century,promising significant benefits for economies and societies. Deep neural networks,like those in Open AI’s ChatGPT, now form the backbone of all modern AI.However, despite AI becoming more common, its core principles are still poorlyunderstood; it remains prone to bias and hallucinations; and it demandsenormous amounts of energy, data, and computational power. In this talk, I willargue that gaining a better understanding of the foundational aspects of modernAI offers a cost-effective way for Australia to take a leading role. The study ofFoundational AI aims to develop methods that are less reliant on big data and highpower consumption. The advancements expected from this research are crucial, ascompanies and governments worldwide race to reduce emissions while expandingthe capabilities of next-generation AI. Investing strategically in Foundational AI willalso strengthen Australia’s capacity for unique sovereign AI capabilities—anessential factor as our country approaches the mid-21st century.

Simon Lucey, Ph.D., is the Director of the Australian Institute for Machine Learning(AIML) at the University of ÌÇÐÄvlog¹ÙÍø, the nation's largest machine learning researchgroup. AIML has secured more than $100 million in funding during hisdirectorship. Professor Lucey previously held key positions at Carnegie MellonUniversity's Robotics Institute, the autonomous vehicle company Argo AI, andCSIRO. He is a scientific advisor on the Temporary AI Expert Committee for theDepartment of Industry, Science and Resources.Professor Lucey has received numerous career awards, including the 2024AmCham Alliance Award for artificial intelligence and an Australian ResearchCouncil Future Fellowship. With 11 patents in computer vision, over 300publications, more than 21,500 citations, and an h-index of 63, his contributions tothe field are widely recognised.His research focuses on computer vision, machine learning, and robotics, drawinginspiration from pioneering AI researchers to uncover computational andmathematical models underlying visual perception.