Next-generation techniques for analysing massive data sets. To process enormous amounts of data, leading computing companies are turning to modern computing frameworks, for which little theory of efficient computational techniques has been developed. This project will resolve key theoretical questions and provide fast techniques for poorly understood pattern recognition and bioinformatics problems.
Fast effective clustering technologies for highly dynamic massive networks. Clustering is a fundamental data mining and analysis task. In an interconnected evolving world, friendships and information flows are modelled as large dynamic networks. Structural clustering and correlation clustering are important and well-studied approaches for static networks; for evolving networks, where links appear and disappear over time, we lack efficient techniques. Anticipated outcomes are new practical cluste ....Fast effective clustering technologies for highly dynamic massive networks. Clustering is a fundamental data mining and analysis task. In an interconnected evolving world, friendships and information flows are modelled as large dynamic networks. Structural clustering and correlation clustering are important and well-studied approaches for static networks; for evolving networks, where links appear and disappear over time, we lack efficient techniques. Anticipated outcomes are new practical clustering algorithms for dynamic networks – with performance guarantees of efficiency and clustering quality – and prototype software, guiding us to pick a good clustering. Expected benefits include better understanding of spread in evolving social networks, accelerating the software testing cycle, and improved topic detection.Read moreRead less
Experimental runtime complexity analysis of logic programs. While declarative languages improve programmer productivity, they make it harder for programmers to understand the performance of their code. We will build a tool that will use profiling data and program analysis to allow programmers to predict the running time of their programs.
Developing Interpretable Machine Learning Models For Clinical Imaging And Single-cell Genomics
Funder
National Health and Medical Research Council
Funding Amount
$1,312,250.00
Summary
Machine learning methods will be vital to make best use of the deluge of data generated by high-throughput technologies in biomedical science. To get the most out of these models, however, we need to be able to unpack the 'black box'. I will use curated clinical and public research data to benchmark and develop interpretable deep learning models and software tools. These models will be used for breast cancer screening programs and for analysis of complex, large-scale single-cell genomics data.
Attribution of Machine-generated Code for Accountability. Machine-generated (or neural) code is usually produced by AI tools to speed up software development. However, such codes have recently raised serious security and privacy concerns. This project aims to attribute these codes to their generative models for accountability purposes. In the process, a series of new techniques are developed to differentiate between the codes generated by different models. The outcomes include analysis of neural ....Attribution of Machine-generated Code for Accountability. Machine-generated (or neural) code is usually produced by AI tools to speed up software development. However, such codes have recently raised serious security and privacy concerns. This project aims to attribute these codes to their generative models for accountability purposes. In the process, a series of new techniques are developed to differentiate between the codes generated by different models. The outcomes include analysis of neural code fingerprints, classification of neural codes, and theories to verify the correctness of code attribution. These will provide significant benefits, ranging from copyright protection to privacy preservation. This project is timely since currently the software community is pervasively using neural codes.Read moreRead less
Fast, practical and effective algorithms for clustering with advice. To maintain a safe and healthy society, government and industry need high quality immunization and national security databases. Since we cannot afford to have duplicate, incomplete and conflicting records that refer to the same person, we unify them by identifying clusters of related records.
In the emerging field of functional genomics, diagnosis of certain diseases is enhanced by determining which genes act together. Diffe ....Fast, practical and effective algorithms for clustering with advice. To maintain a safe and healthy society, government and industry need high quality immunization and national security databases. Since we cannot afford to have duplicate, incomplete and conflicting records that refer to the same person, we unify them by identifying clusters of related records.
In the emerging field of functional genomics, diagnosis of certain diseases is enhanced by determining which genes act together. Different experimental runs might result in different clusterings of genes: we need one consensus clustering that summarizes the experimental outcomes.
Cleaning databases and combining clusterings by hand would require vast amounts of time. This project will result in faster and more accurate computational procedures.Read moreRead less
Unlocking Hidden Cancer Drivers Using Transcriptome Data
Funder
National Health and Medical Research Council
Funding Amount
$700,473.00
Summary
New sequencing technologies allow us to get an unbiased look at the molecular signalling in a tumour. However this information is very complex and need specialised methods in statistic and computation in order to make new discoveries. Here will will develop analysis methods to find novel transcriptional variants in cancer and then test them in the lab in order to understand if our discoveries are responsible for causing cancer.
Algorithms and computation in four-dimensional topology. This project will establish Australia as a world leader in computational topology, particularly in the all-important areas of topology in three and four dimensions. In four dimensions this work will be truly groundbreaking; until now the field has seen little development due to the complexity of the algorithms and computations required, and the applicant is in the unique position of having the necessary tools to make significant progress ....Algorithms and computation in four-dimensional topology. This project will establish Australia as a world leader in computational topology, particularly in the all-important areas of topology in three and four dimensions. In four dimensions this work will be truly groundbreaking; until now the field has seen little development due to the complexity of the algorithms and computations required, and the applicant is in the unique position of having the necessary tools to make significant progress in a feasible time frame. In three dimensions this project will strengthen the distinguished computational topology community in Melbourne, led by pioneers such as Rubinstein, Goodman, Hodgson as well as the applicant himself.Read moreRead less
Efficient Synchronisation of Large Repositories. Accuracy and maintenance of vast quantities of data are essential for any modern society. The economy, health institutes and industries, and our defence and legal systems rely on having data being distributed widely and securely, and on queries being answered accurately and quickly. Complete synchronisation of databases is often impossible due to the limitations of internet bandwidth. Better compression techniques have the potential to allow crit ....Efficient Synchronisation of Large Repositories. Accuracy and maintenance of vast quantities of data are essential for any modern society. The economy, health institutes and industries, and our defence and legal systems rely on having data being distributed widely and securely, and on queries being answered accurately and quickly. Complete synchronisation of databases is often impossible due to the limitations of internet bandwidth. Better compression techniques have the potential to allow critical data to be distributed much more efficiently; we anticipate in some applications that the size of a compressed file could be reduced tenfold or more compared to previous best methods, leading to dramatic savings.Read moreRead less
Quantum decoherence: A game-theoretic perspective. Algorithms based on quantum computation have the ability to significantly speed up information processing compared to standard computers. The increase in computational power can have enormous impact on humankind and this project will help maintain Australia's position in the global forefront of this effort.This project focuses on the thoeretical foundations of quantum computation and complements the efforts of several groups in Australia collabo ....Quantum decoherence: A game-theoretic perspective. Algorithms based on quantum computation have the ability to significantly speed up information processing compared to standard computers. The increase in computational power can have enormous impact on humankind and this project will help maintain Australia's position in the global forefront of this effort.This project focuses on the thoeretical foundations of quantum computation and complements the efforts of several groups in Australia collaborating on the experimental design of quantum computers. The project will increase the fundamental understanding of how quantum information is processed in the presence of noise, which is necessary for the successful operation of quantum computers. Read moreRead less