We’re not building robots to take the place of humans, but creating technology that will work alongside people to help make more informed and better decisions.
Focus areas of A²I²
- Bayesian optimisation for experimental design
- Large-scale surveillance
- Advanced video understanding
- Software engineering for AI
- Reinforcement learning
- Deep learning models for health care
- Value-based machine learning
- Pattern recognition
- Personalised medicine
- Social media analysis for improved mental health
Toby Playpad is a novel, cost-effective and flexible framework for early intervention, facilitating therapy for children with autism. Our fundamental research resulted in the release of an iPad app in September 2012 that is now used by more than 7000 people worldwide. A uniquely adaptive early intervention program for children with autism, TOBY Playpad monitors each child’s performance on tasks from a complex syllabus and adjusts lesson delivery to each and every child. Parent feedback tells us that, thanks to TOBY, their children are learning new social, cognitive and sensory skills. The technology won the International Education – Research Engagement Award (Victorian Government) in 2014 and the Geelong Chamber of Commerce Smart Technology Award in 2012. It has also triggered two clinical trials, led by the Telethon Child Health Institute.
Intelligent IoT platform
To assist the elderly and those with disabilities to live independently for in their own homes, A²I² has developed an Internet of Things-based solution that uses low cost sensors placed around the home to passively learn an individual's regular routine. Understanding this routine enables the platform to identify behavioural anomalies like a fall in the shower and raise an alert.
Leveraging our long collaboration with the National Trauma Research Institute, we’re improving the treatment that trauma patients receive from the time they’re picked up by the ambulance through to discharge. We worked with The Alfred, one of Australia’s busiest emergency and trauma centres to build a decision support system that helps reduce errors of omission in the first half an hour of a patient’s arrival. Independent analysis of the system showed it cut errors of omission by 22% (that’s the equivalent of saving 5 or 6 lives per year), reduced time in ICU and produced better patient outcomes.
A2I2 is working with Defence on armour design, next generation training systems, operations modelling and the construction of virtual environments.
SeeUV is is an augmented reality educational application developed teach young Australians about the danger of ultraviolet radiation. As ultraviolet radiation can’t be detected by human senses, A2I2 collaborated with Cancer Council Victoria to design a solution that simulates the potential damage that UV can cause. To make UV more tangible, the seeUV uses an augmented reality filter to create a virtual representation of ultraviolet radiation. These beams are a representation based on the UV index provided by the Bureau of Meteorology, and rotate based on the sun's location.
Detecting security threats
We've discovered new techniques that detect potential security threats in large data sets, such as those collected by surveillance cameras. Our algorithms power innovative anomaly detection software called iCetana, which is now used in video surveillance worldwide. The software uses compressed sensing concepts to enable simultaneous surveillance of many cameras deployed in diverse settings. iCetana received the Broadband Innovation Award, Tech23 in 2010 and the 2011 WA Innovator of the Year award.
Applied Artificial Intelligence Institute