Deep visual understanding: learning to see in an unruly world. Deep Learning has achieved incredible success at an astonishing variety of Computer Vision tasks recently. This project aims to convey this success into the challenging domain of high-level image-based reasoning. It will extend deep learning to achieve flexible semantic reasoning about the content of images based on information gleaned from the huge volumes of data available on the Internet. The project expects to overcome one of the ....Deep visual understanding: learning to see in an unruly world. Deep Learning has achieved incredible success at an astonishing variety of Computer Vision tasks recently. This project aims to convey this success into the challenging domain of high-level image-based reasoning. It will extend deep learning to achieve flexible semantic reasoning about the content of images based on information gleaned from the huge volumes of data available on the Internet. The project expects to overcome one of the primary limitations of deep learning and will greatly increase its practical application to a range of industrial, cultural or health settings.Read moreRead less
Action selection in insects: how a microbrain knows what to do. Identifying what to do demands integrating sensory information with our current physiological state and memory of past experience to select the best possible action. This is the action selection problem. Our project aims to discover how tiny insect brains solve this fundamental problem. The project combines neural recordings from animals exploring virtual reality, behavioural analyses and computational modelling. The expected outco ....Action selection in insects: how a microbrain knows what to do. Identifying what to do demands integrating sensory information with our current physiological state and memory of past experience to select the best possible action. This is the action selection problem. Our project aims to discover how tiny insect brains solve this fundamental problem. The project combines neural recordings from animals exploring virtual reality, behavioural analyses and computational modelling. The expected outcome is a new understanding of the brain as an effective behavioural control system. This will benefit systems and comparative neuroscience. Our findings may also inspire solutions for robotic systems that must operate autonomously in remote and challenging environments such as disaster relief or exploration.Read moreRead less
Making Meta-learning Generalised . This project aims to develop novel machine learning techniques, termed generalised meta-learning, to make machines better utilise past experience to solve new tasks with few data. It expects to reduce the undesirable dependence of current machine learning on labelled data and significantly expand its application scope. Expected outcomes of the project consist of new theoretical results on meta-learning and a set of innovative algorithms that can support the bui ....Making Meta-learning Generalised . This project aims to develop novel machine learning techniques, termed generalised meta-learning, to make machines better utilise past experience to solve new tasks with few data. It expects to reduce the undesirable dependence of current machine learning on labelled data and significantly expand its application scope. Expected outcomes of the project consist of new theoretical results on meta-learning and a set of innovative algorithms that can support the building of next generation of computer vision systems to work in open and dynamic environments. This should be able to produce solid benefits to the science, society, and economy of Australian via the application of these advanced intelligent systems.Read moreRead less
Things don’t always go better with Coke. This project aims to test whether soft drink use is governed partly by automatic processes (cognitive biases) that operate largely outside of conscious control. In so doing, the project expects to generate a new conceptual understanding of the mechanisms that drive the overconsumption of soft drinks. Expected outcomes include theoretical innovation, new research methodologies, and accessible cost-effective technologies for reducing excessive sugar intake ....Things don’t always go better with Coke. This project aims to test whether soft drink use is governed partly by automatic processes (cognitive biases) that operate largely outside of conscious control. In so doing, the project expects to generate a new conceptual understanding of the mechanisms that drive the overconsumption of soft drinks. Expected outcomes include theoretical innovation, new research methodologies, and accessible cost-effective technologies for reducing excessive sugar intake from soft drinks, in line with recent World Health Organization guidelines. These outcomes will contribute to combatting obesity and tooth decay.Read moreRead less
Life or death decisions: making fast, accurate choices in a complex world. This project aims to understand how hoverflies and honey bees, with tiny brains and sensory systems, excel at making fast and accurate decisions while on the fly in a complex world. The project will combine brain recordings with flight analyses and computational modelling to generate new knowledge on how animals may utilize movements to simplify information sampling. Expected outcomes are a novel, comprehensive understand ....Life or death decisions: making fast, accurate choices in a complex world. This project aims to understand how hoverflies and honey bees, with tiny brains and sensory systems, excel at making fast and accurate decisions while on the fly in a complex world. The project will combine brain recordings with flight analyses and computational modelling to generate new knowledge on how animals may utilize movements to simplify information sampling. Expected outcomes are a novel, comprehensive understanding of how animal movements could enhance decision speed and accuracy. This should provide substantial benefits for neuroscience, and for enhancing performance of autonomous robotic systems operating in challenging environments, such as disaster relief, mining and remote exploration. Read moreRead less
Comparative analysis of sensor noise for target detection in dragonfly eyes. Dragonflies hunt tiny prey in the low-light conditions of late dusk, a signal-to-noise problem that challenges any engineered system. Using a comparative approach across dragonfly species, we aim to use novel optical and physiological measures to determine how sensors with noise underlie target-detection, in varying scene brightness. The project outcomes will be a comparative characterisation of signal-to-noise measures ....Comparative analysis of sensor noise for target detection in dragonfly eyes. Dragonflies hunt tiny prey in the low-light conditions of late dusk, a signal-to-noise problem that challenges any engineered system. Using a comparative approach across dragonfly species, we aim to use novel optical and physiological measures to determine how sensors with noise underlie target-detection, in varying scene brightness. The project outcomes will be a comparative characterisation of signal-to-noise measures of dragonfly eye optics (including eye size) and early sensory neurons. We will match detection thresholds with downstream target-detecting neurons and dragonfly behaviour. This will provide insight into signal detection, which is a ubiquitous problem across information processing, computer vision and autonomous systems.Read moreRead less
New Paradigms for Robust Fitting: Kernelisation and Polyhedral Search. Outliers inevitably exist in visual data due to imperfect data acquisition or preprocessing. To enable computer vision applications that can perform reliably, robust fitting algorithms are necessary to counter the biasing influence of outliers. However, current robust algorithms are unsatisfactory: they are unreliable (due to using randomisation) or too computationally costly (due to using exhaustive search). This project wil ....New Paradigms for Robust Fitting: Kernelisation and Polyhedral Search. Outliers inevitably exist in visual data due to imperfect data acquisition or preprocessing. To enable computer vision applications that can perform reliably, robust fitting algorithms are necessary to counter the biasing influence of outliers. However, current robust algorithms are unsatisfactory: they are unreliable (due to using randomisation) or too computationally costly (due to using exhaustive search). This project will develop new robust algorithms to mitigate these shortcomings. It will do so by investigating two new paradigms of kernelisation and polyhedral search, which offer unprecedented theoretical insights into the problem. The outcomes will contribute towards computer vision applications that are more practical and reliable.Read moreRead less
Fairness aware data mining for discrimination free decision-making. This project aims to develop data mining methods to detect algorithmic discriminations and to build fair decision models. It expects to provide techniques for regulatory organisations to detect discriminations in algorithmic decisions, and for various companies and organisations to build fair decision systems. Expected outcomes are novel and accurate methods for discrimination detection, practical and versatile techniques for fa ....Fairness aware data mining for discrimination free decision-making. This project aims to develop data mining methods to detect algorithmic discriminations and to build fair decision models. It expects to provide techniques for regulatory organisations to detect discriminations in algorithmic decisions, and for various companies and organisations to build fair decision systems. Expected outcomes are novel and accurate methods for discrimination detection, practical and versatile techniques for fair decision model building, and improved understanding of the relationships between privacy preservation and discrimination prevention to enable new techniques to achieve both goals. The developed techniques enable society to tackle ethical challenges in the big data era where many decisions are analytics based. Read moreRead less
Evolutionary diversity optimisation. This project aims to build up and establish the area of evolutionary diversity optimisation. The project will cover the design and application of evolutionary diversity optimisation methods to complex problems of significance and high national economic benefit and build up the theoretical foundations of these methods. The project is expected benefit decision makers by providing them a diverse set of high quality alternatives to choose from. This project will ....Evolutionary diversity optimisation. This project aims to build up and establish the area of evolutionary diversity optimisation. The project will cover the design and application of evolutionary diversity optimisation methods to complex problems of significance and high national economic benefit and build up the theoretical foundations of these methods. The project is expected benefit decision makers by providing them a diverse set of high quality alternatives to choose from. This project will allow them to make highly informed decisions and lead to more reliable solutions for optimisation problems, in areas of high economic impact such as manufacturing and supply chain management.Read moreRead less
Learning Robotic Navigation and Interaction from Object-based Semantic Maps. Our project aims to develop new learning algorithms that enable robots to perform high-complexity tasks that are currently impossible. Compared to existing methods that rely on low-level sensor data, we aim to achieve this by learning from a high-level graph representation of the environment that captures semantics, affordances, and geometry. The outcome would be robots capable of using human instructions to efficiently ....Learning Robotic Navigation and Interaction from Object-based Semantic Maps. Our project aims to develop new learning algorithms that enable robots to perform high-complexity tasks that are currently impossible. Compared to existing methods that rely on low-level sensor data, we aim to achieve this by learning from a high-level graph representation of the environment that captures semantics, affordances, and geometry. The outcome would be robots capable of using human instructions to efficiently learn complex interaction and navigation behaviours that transfer to unseen environments. Our research should benefit new applications in domains of economic and societal importance that are currently too complex, unsafe, and uncertain for robot assistants, such as aged care, advanced manufacturing and domestic robotics.Read moreRead less