Razi said the team won the ONI and Department of Defence National Security Science and Technology Centre grant because a new type of machine intelligence that could “learn throughout its lifetime” was needed.
Such intelligence would improve machine learning for technology including self-driving cars, autonomous drones and delivery robots, he said.
“This new technology capability in the future may eventually surpass the performance of existing, purely silicon-based hardware.
“The outcomes of such research would have significant implications across multiple fields such as – but not limited to – planning, robotics, advanced automation, brain-machine interfaces, and drug discovery, giving Australia a significant strategic advantage.”
Brains are good at lifelong learning, which is needed to gain new skills, adapt to change, and apply existing knowledge to new tasks, while artificial intelligence suffers from what researchers call “catastrophic forgetting”. AI forgets information from previous tasks when it starts new ones.
The DishBrain research aims to understand the biological mechanisms behind ongoing learning.
“We will be using this [national intelligence and security discovery research] grant to develop better AI machines that replicate the learning capacity of these biological neural networks,” Razi said.
“This will help us scale up the hardware and methods capacity to the point where they become a viable replacement for in-silico computing [using simulations].”
The news comes as AI leaders call on the government to recognise “the potential for catastrophic or existential risks from AI”.
The organisation Australians For AI Safety has written a letter to industry, science and technology minister, Ed Husic, signed by academics and industry heads.