Estimating the Topology of Low-Dimensional Data Using Deep Neural Networks. This project will expand on the superhuman visual capabilities of deep neural networks to allow us to analyse the topology of 3- and 4-dimensional manifolds. While these spaces still count as low-dimensional, 4-dimensional manifolds typically are beyond human visual comprehension. The topology of a manifold describes its essential properties such as the number of connected components, holes, tunnels and cavities of vario ....Estimating the Topology of Low-Dimensional Data Using Deep Neural Networks. This project will expand on the superhuman visual capabilities of deep neural networks to allow us to analyse the topology of 3- and 4-dimensional manifolds. While these spaces still count as low-dimensional, 4-dimensional manifolds typically are beyond human visual comprehension. The topology of a manifold describes its essential properties such as the number of connected components, holes, tunnels and cavities of various dimensions. Traditional methods from computational topology fail in large practical applications due to computational restrictions. We propose an approximation that overcomes previous limitations and can open new doors to data analysis in material science, medical imaging, dynamical systems and other applications.
Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE200100063
Funder
Australian Research Council
Funding Amount
$394,398.00
Summary
Nonmonotone Algorithms in Operator Splitting, Optimisation and Data Science. This project aims to develop the mathematical foundations for the analysis and development of optimisation algorithms used in data science. Despite their now ubiquitous use, machine learning software packages routinely rely on a number of algorithms from mathematical optimisation which are not properly understood. By moving beyond the traditional realms of Fejér monotone algorithms, this project expects to develop the m ....Nonmonotone Algorithms in Operator Splitting, Optimisation and Data Science. This project aims to develop the mathematical foundations for the analysis and development of optimisation algorithms used in data science. Despite their now ubiquitous use, machine learning software packages routinely rely on a number of algorithms from mathematical optimisation which are not properly understood. By moving beyond the traditional realms of Fejér monotone algorithms, this project expects to develop the mathematical theory required to rigorously justify the use of such algorithms and thereby ensure the integrity of the decision tools they produce. This mathematical framework is also expected to produce new algorithms for optimisation which benefit consumers of data science such as the health-care and cybersecurity sectors.Read moreRead less
Hybrid methods with decomposition for large scale optimization. This project aims to develop advanced approaches for solving large scale real-world optimisation problems that are expensive to evaluate, and difficult to formulate, involving thousands of variables and constraints. The project will make novel contributions to improving state-of-the-art large scale optimisation algorithms in terms of scalability, effectiveness, and efficiency for real-world problem solving. The outcomes of this proj ....Hybrid methods with decomposition for large scale optimization. This project aims to develop advanced approaches for solving large scale real-world optimisation problems that are expensive to evaluate, and difficult to formulate, involving thousands of variables and constraints. The project will make novel contributions to improving state-of-the-art large scale optimisation algorithms in terms of scalability, effectiveness, and efficiency for real-world problem solving. The outcomes of this project will bring about greater understanding of real-world large scale optimisation, and deliver practical solutions to these problems.Read moreRead less
Democratisation of Deep Learning: Neural Architecture Search at Low Cost. The need to manually design Deep Learning-based Neural Networks (DNNs) limits their usage to AI experts and hinders the exploitation of their true potential more broadly, e.g., in farming, humanities. We aim to replace this tedious process through novel AI methods capable of generating DNNs that can perform significantly better and at a lower computational cost than manually designed DNNs. We further expand this idea to so ....Democratisation of Deep Learning: Neural Architecture Search at Low Cost. The need to manually design Deep Learning-based Neural Networks (DNNs) limits their usage to AI experts and hinders the exploitation of their true potential more broadly, e.g., in farming, humanities. We aim to replace this tedious process through novel AI methods capable of generating DNNs that can perform significantly better and at a lower computational cost than manually designed DNNs. We further expand this idea to solve complex real-world problems with both labelled and unlabelled data found in various applications including energy and climate change. The expected outcomes include the novel AI methods, highly trained AI researchers and a number of critical applications that will bring significant benefits to Australia and the world.Read moreRead less
Machine education for trusted multi-skilled evolutionary learners . Transforming data assets into organisational knowledge assets sits in the hands of a few, highly specialised, data scientists. The aim of this research is to design educational instruments to support non-experts to teach artificial intelligence (AI) systems in a similar way to educating human teachers to teach human learners. The significance of the project lies in affording the wider smart, but not necessarily AI expert, commun ....Machine education for trusted multi-skilled evolutionary learners . Transforming data assets into organisational knowledge assets sits in the hands of a few, highly specialised, data scientists. The aim of this research is to design educational instruments to support non-experts to teach artificial intelligence (AI) systems in a similar way to educating human teachers to teach human learners. The significance of the project lies in affording the wider smart, but not necessarily AI expert, community the ability to contribute to growing our knowledge-based society in a safe, transparent and trustworthy manner. Outcomes will include innovative instruments to teach machines, novel knowledge creation, trusted and transparent AI systems, and a new generation of human teachers specialised in educating AI systems.Read moreRead less
Memetic algorithms for multiobjective optimisation problems in bioinformatics. Many questions of paramount importance in life sciences can be formulated as optimisation problems but using just a single criterion can be misleading. This project will address this problem using multiobjective optimisation and leveraging Australia's investment in supercomputing with algorithms that mimic evolutionary processes in silico.
Memetic algorithms and adaptive memory metaheuristics for large scale combinatorial optimisation problems arising in biomarker discovery. Despite modern supercomputers, the world depends on combinatorial optimisation, the branch of mathematics and computer science that involves finding optimal solutions when it is impossible to enumerate all solutions. We bring complementary skills to address the core set of the five most challenging problems arising from novel biotechnologies.
Unleashing the power of a supernetwork-driven approach for bioinformatics. Supernetworks are built “above and beyond” existing networks. In bioinformatics they arise from the integration of a set of networks with different types of nodes and edges. While large companies and governments have already understood the importance of decision problems in supernetworks, the power of this perspective has not yet been exploited in the life sciences. Using supercomputer-based approaches together with memet ....Unleashing the power of a supernetwork-driven approach for bioinformatics. Supernetworks are built “above and beyond” existing networks. In bioinformatics they arise from the integration of a set of networks with different types of nodes and edges. While large companies and governments have already understood the importance of decision problems in supernetworks, the power of this perspective has not yet been exploited in the life sciences. Using supercomputer-based approaches together with memetic algorithms, this project will address the first key areas that will lead to smart information use of existing networks in distributed databases. This project will deliver the next generation of algorithms for network alignment, identification of connected-cohesive subnetworks and embedding large graphs in multi-planar structures.Read moreRead less
Challenging systems to discover vulnerabilities using computational red teaming. Computational red teaming concerns the design of computational models to role play intelligent adversaries. These adversaries who are determined to exploit a system rely on creative thinking to discover system-level vulnerabilities by challenging system design, implementation or operations. This project closes a gap in the risk assessment literature by designing automated computational red teaming methods to discove ....Challenging systems to discover vulnerabilities using computational red teaming. Computational red teaming concerns the design of computational models to role play intelligent adversaries. These adversaries who are determined to exploit a system rely on creative thinking to discover system-level vulnerabilities by challenging system design, implementation or operations. This project closes a gap in the risk assessment literature by designing automated computational red teaming methods to discover vulnerabilities associated with intentional risks. Scientific outcomes include novel automated skill assessment algorithms and new search techniques to exploit assumptions that could have been overlooked otherwise. Practical outcomes include robust risk assessment tools, strong research training, and high impact publications.Read moreRead less
Evolutionary computation for expensive bilevel multiobjective problems. This project aims to develop an evolutionary computation framework to solve computationally expensive bilevel multiobjective problems. The research is fundamental in nature and will address key open challenges in solving such problems, including hierarchical decision-making, multiple performance criteria, uncertainties and computational expense. The proposed research has applications in diverse domains such as environmental ....Evolutionary computation for expensive bilevel multiobjective problems. This project aims to develop an evolutionary computation framework to solve computationally expensive bilevel multiobjective problems. The research is fundamental in nature and will address key open challenges in solving such problems, including hierarchical decision-making, multiple performance criteria, uncertainties and computational expense. The proposed research has applications in diverse domains such as environmental policy formulation, network design, engineering, defence and cybersecurity; offering significant benefits to the researchers and practitioners in these fields. In addition to research outputs, it will strengthen international collaboration and build research capacity to put Australia at the forefront of this research.
Read moreRead less