Partially Observable MDPs, Monte Carlo Methods, and Sustainable Fisheries. Partially Observable Markov Decision Processes (POMDPs) provide a general mathematical framework for sequential decision making under uncertainty. However, solving POMDPs effectively under realistic assumptions remains a challenging problem. This project aims to develop new efficient Monte Carlo algorithms to significantly advance the application of POMDPs to real-world decision problems involving complex action spaces an ....Partially Observable MDPs, Monte Carlo Methods, and Sustainable Fisheries. Partially Observable Markov Decision Processes (POMDPs) provide a general mathematical framework for sequential decision making under uncertainty. However, solving POMDPs effectively under realistic assumptions remains a challenging problem. This project aims to develop new efficient Monte Carlo algorithms to significantly advance the application of POMDPs to real-world decision problems involving complex action spaces and system dynamics. Both theoretical and algorithmic approaches will be applied to sustainable fishery management --- an important problem for Australia and an ideal context for POMDPs. The project will advance research in artificial intelligence, dynamical systems, and fishery operations, and benefit the national economy.Read moreRead less
Techniques to use stereo vision for improving person identification systems based on face recognition. The broad aim of this project is to use three-dimensional information available by processing images from stereo cameras in order to bridge the gap between constrained face recognition systems and viable systems that work well under varying illumination, changes in pose and variations in spectacles, facial hair and attire. Such a system will be useful in passenger verification at airports and a ....Techniques to use stereo vision for improving person identification systems based on face recognition. The broad aim of this project is to use three-dimensional information available by processing images from stereo cameras in order to bridge the gap between constrained face recognition systems and viable systems that work well under varying illumination, changes in pose and variations in spectacles, facial hair and attire. Such a system will be useful in passenger verification at airports and as a component of personal identification systems to counter terrorism. The key to successful face location and recognition is an effective combination of all data - range, luminance and colour - and techniques for this will be the discovered outcomes.Read moreRead less
Searching for Maximal Satisfaction. A wide range of practical problems such as scheduling, timetabling, planning and economic forecasting are not only computationally intractable in general, but often involve conflicting constraints that make them unsolvable. These problems can be represented as MaxSAT, the optimisation version of the satisfiability problem (SAT). This project aims to develop novel and efficient algorithms to address the problem of maximal satisfaction. It is proposed that these ....Searching for Maximal Satisfaction. A wide range of practical problems such as scheduling, timetabling, planning and economic forecasting are not only computationally intractable in general, but often involve conflicting constraints that make them unsolvable. These problems can be represented as MaxSAT, the optimisation version of the satisfiability problem (SAT). This project aims to develop novel and efficient algorithms to address the problem of maximal satisfaction. It is proposed that these algorithms will be implemented within prototype MaxSAT solver systems, which will be experimentally evaluated on large-sized real world optimisation problems of high economic and societal significance. These solvers are expected to also compete in the industrial track of the international SAT solving competitions.Read moreRead less
An automated 3D model-based object recognition system. A novel, practical 3D vision system is proposed as a platform for fundamental applied research in 3D data acquisition, object modelling and object recognition. The significance of the vision system lies in the advancement of knowledge in three key areas of computer vision, registration, recognition and error propagation. The result is a system capable of sensing, modelling and identifying arbitrarily shaped free-form objects in a scene, an a ....An automated 3D model-based object recognition system. A novel, practical 3D vision system is proposed as a platform for fundamental applied research in 3D data acquisition, object modelling and object recognition. The significance of the vision system lies in the advancement of knowledge in three key areas of computer vision, registration, recognition and error propagation. The result is a system capable of sensing, modelling and identifying arbitrarily shaped free-form objects in a scene, an attribute lacking in current systems. Such a system can provide substantial economic benefits to industrial procedures such as grasp planning and quality control.Read moreRead less
Advanced Monte Carlo Methods for Spatial Processes. The modeling and analysis of spatial data relies more and more on sophisticated Monte Carlo simulation methods. However, with the growing complexity of today's spatial data, traditional Monte Carlo methods increasingly face difficulties in terms of speed and accuracy. The aim of this project is to develop new theory and applications at the interface of Monte Carlo methods and spatial statistics, building upon exciting theoretical and computatio ....Advanced Monte Carlo Methods for Spatial Processes. The modeling and analysis of spatial data relies more and more on sophisticated Monte Carlo simulation methods. However, with the growing complexity of today's spatial data, traditional Monte Carlo methods increasingly face difficulties in terms of speed and accuracy. The aim of this project is to develop new theory and applications at the interface of Monte Carlo methods and spatial statistics, building upon exciting theoretical and computational advances in both areas in recent years. The research will stimulate the design of microscopic and macroscopic complex spatial structures with superior properties, such as composite materials, solar cells, telecommunication networks, mining operations, and road systems.Read moreRead less
Automatic Brain Tissue Segmentation in Magnetic Resonance Images based on Knowledge-guided Constrained Clustering. Accurate volumetric measurement of brain tissues is of critical importance in the study of many brain disorders, disease diagnosis, disease progression tracking and treatment monitoring. The study in this research will result in the development of a powerful computational technique that allows automatic volumetric measurement and analysis of brain tissues. The software developed in ....Automatic Brain Tissue Segmentation in Magnetic Resonance Images based on Knowledge-guided Constrained Clustering. Accurate volumetric measurement of brain tissues is of critical importance in the study of many brain disorders, disease diagnosis, disease progression tracking and treatment monitoring. The study in this research will result in the development of a powerful computational technique that allows automatic volumetric measurement and analysis of brain tissues. The software developed in this project will expedite early clinical diagnosis and treatment of neural diseases for patients, hence saving life and reducing health cost both at the personal and the national level. Read moreRead less
High resolution single particle analysis of biological macromolecules. One of the great challenges of cell biology is to increase the rate of atomic resolution structure determination, particularly of membrane proteins and macromolecular assemblies. The current rate-limiting step is high quality crystal production. Our goal is to prove that protein structures can be determined to atomic resolution by single-particle analysis. 3D structures will be produced by computationally aligning high-resolu ....High resolution single particle analysis of biological macromolecules. One of the great challenges of cell biology is to increase the rate of atomic resolution structure determination, particularly of membrane proteins and macromolecular assemblies. The current rate-limiting step is high quality crystal production. Our goal is to prove that protein structures can be determined to atomic resolution by single-particle analysis. 3D structures will be produced by computationally aligning high-resolution electron microscope images of individual, randomly oriented molecules. The importance of this project is highlighted by the fact over 120,000 protein sequences are already databased, a number set to increase rapidly as new genome sequencing projects are completed.
Read moreRead less
Dynamics of Causal Knowledge. We operate in complex dynamic environments including highly sensitive and safety-critical situations such as medical emergencies, disaster management and air-traffic control systems. Our knowledge of what causes what plays a pivotal role in making correct decisions in such situations. To ensure robustness and sound behaviour of the underlying causal knowledge systems, their designs and implementations must be formally well grounded. This is an important but difficul ....Dynamics of Causal Knowledge. We operate in complex dynamic environments including highly sensitive and safety-critical situations such as medical emergencies, disaster management and air-traffic control systems. Our knowledge of what causes what plays a pivotal role in making correct decisions in such situations. To ensure robustness and sound behaviour of the underlying causal knowledge systems, their designs and implementations must be formally well grounded. This is an important but difficult challenge. This project aims to systematically develop a logic-based framework to adequately capture and reason about evolving causal knowledge. This research is expected to form the basis for smart decision making, and be evaluated on practical applications.Read moreRead less
Improved detection and characterisation of breast cancer using magnetic resonance imaging, and novel image analysis and pattern recognition techniques. Breast cancer is a leading cause of death in Australian women. With no clear cause, one mainstay of management has been early detection. Newer medical imaging technologies such as magnetic resonance imaging require complex analysis to achieve their full benefit. Should the computationally demanding analyses of these images provide more sensitive ....Improved detection and characterisation of breast cancer using magnetic resonance imaging, and novel image analysis and pattern recognition techniques. Breast cancer is a leading cause of death in Australian women. With no clear cause, one mainstay of management has been early detection. Newer medical imaging technologies such as magnetic resonance imaging require complex analysis to achieve their full benefit. Should the computationally demanding analyses of these images provide more sensitive and specific detection of early cancers, the potential reductions in morbidity and mortality from breast cancer will be of immense value. Successful implementation of the proposed project will further enhance Australia's position as a world leader in biomedical research and application of computational technologies to health problems.Read moreRead less
Searching for near-exact protein models. This project aims to develop novel and efficient heuristic-based algorithms leading to near accurate protein tertiary structure models. Knowledge about protein structures is fundamental to our understanding of living systems. The progress on experimental determination of these structures has been extremely limited and remains an open challenge in molecular biology. Computational prediction of protein structures from sequences is emerging as a promising ap ....Searching for near-exact protein models. This project aims to develop novel and efficient heuristic-based algorithms leading to near accurate protein tertiary structure models. Knowledge about protein structures is fundamental to our understanding of living systems. The progress on experimental determination of these structures has been extremely limited and remains an open challenge in molecular biology. Computational prediction of protein structures from sequences is emerging as a promising approach, but its accuracy is far from satisfactory. The software systems developed in this project will be used in structural identification of target proteins in drug design. This will make drug design process more efficient, saving time and cost, potentially saving lives.Read moreRead less