The Machine Intelligence Lab conducts world class research to harness the transition from a process-defined world to a data-driven one by creating and developing future AI technologies and techniques that will have transformational effects across the economy and society. We contribute to the methodology and technology of diverse areas, such as computer vision, logistics, agriculture, environmental and resource management, health, defence and IoT.

Management of financial risk is critical for organisations, particularly those relying on risk exposure for competitive advantage. The task, however, has become increasingly complex due to the fast-changing risk landscape. We are leveraging the latest advances in artificial intelligence (AI) to develop better models for identifying and measuring risks. It will produce novel machine learning algorithms that can assist decision-makers to turn the risk into business performance.

We are using advanced analytics techniques to explore sports, in this case Netball, team and player statistics primarily obtained from an athlete management system. The idea is to find significant team and player performance patterns, and support the sports’ decision making process from both the athlete selection and strategic game planning perspectives.

The concept of Al-enhanced instrumentation is based on the notion that one or more original sources of data can be post-processed and enhanced using machine learning to bring out added detail, remove noise or improve sensitivity. We are working on the use of deep learning methods to develop, model and simulate Al-enhanced instrumentation. This is since deep networks have been very successfully used in recognition and physics-based vision, where the image formation process is used to constrain the loss function used to train the network.

We are working to deliver underwater, proximal and remote optical sensing systems and solutions tailored specifically to the Australian aquatic ecosystem needs, but with global applicability. To do this, we are using deep networks, statistical and syntactical pattern recognition techniques for recognition and classification of underwater fauna and flora. We are also developing sensor fusion technologies for combining spatial and optical cues into the mapping of underwater environments.