Exploiting Geometries of Learning for Fast, Adaptive and Robust AI. This project aims to uniquely exploit geometric manifolds in deep learning to advance the frontier of Artificial Intelligence (AI) research and applications in cybersecurity and general cognitive tasks. It expects to develop new theories, algorithms, tools, and technologies for machine learning systems that are fast, adaptive, lifelong and robust, even with limited supervision. Expected outcomes will enhance Australia's capabili ....Exploiting Geometries of Learning for Fast, Adaptive and Robust AI. This project aims to uniquely exploit geometric manifolds in deep learning to advance the frontier of Artificial Intelligence (AI) research and applications in cybersecurity and general cognitive tasks. It expects to develop new theories, algorithms, tools, and technologies for machine learning systems that are fast, adaptive, lifelong and robust, even with limited supervision. Expected outcomes will enhance Australia's capability and competitiveness in AI, and deliver robust and trustworthy learning technology. The project should provide significant benefits not only in advancing scientific and translational knowledge but also in accelerating AI innovations, safeguarding cyberspace, and reducing the burden on defence expenses in Australia.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE240100816
Funder
Australian Research Council
Funding Amount
$444,000.00
Summary
Probing dark energy with the largest 3D Map of the Universe. Dark Energy is one of the most profound mysteries of modern physics. It makes up about 70 percent of the Universe, but no compelling theory can explain its nature. This project aims to measure the properties of Dark Energy with unprecedented accuracy: an order of magnitude better than the state of the art. It aims to accomplish this by extracting information from the largest 3D map of the cosmos, built with the optical spectra of 35 mi ....Probing dark energy with the largest 3D Map of the Universe. Dark Energy is one of the most profound mysteries of modern physics. It makes up about 70 percent of the Universe, but no compelling theory can explain its nature. This project aims to measure the properties of Dark Energy with unprecedented accuracy: an order of magnitude better than the state of the art. It aims to accomplish this by extracting information from the largest 3D map of the cosmos, built with the optical spectra of 35 million galaxies, observed by the Dark Energy Spectroscopic Instrument. This project will foster Australia's historic leadership and investments in galaxy surveys via unique international partnerships, and produce cutting-edge tools for big data analyses with important applications in a wide range of industries.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE240101089
Funder
Australian Research Council
Funding Amount
$436,847.00
Summary
Trustworthy Hypothesis Transfer Learning. It is urgent to develop a new hypothesis transfer learning scheme that can overcome potential risks when finetuning unreliable large-scale pre-trained models. This project aims to develop an advanced and reliable scheme of hypothesis transfer learning, called Trustworthy Hypothesis Transfer Learning (TrustHTL). A new theoretically guaranteed heterogeneous hypothesis transfer learning framework will be developed to handle heterogeneous situations; a metho ....Trustworthy Hypothesis Transfer Learning. It is urgent to develop a new hypothesis transfer learning scheme that can overcome potential risks when finetuning unreliable large-scale pre-trained models. This project aims to develop an advanced and reliable scheme of hypothesis transfer learning, called Trustworthy Hypothesis Transfer Learning (TrustHTL). A new theoretically guaranteed heterogeneous hypothesis transfer learning framework will be developed to handle heterogeneous situations; a methodology to disinherit risks of pre-trained models and a new fuzzy relation based distributional discrepancy in heterogeneous transfer learning scenarios. The outcomes should significantly improve the reliability of machine learning with benefits for safety learning in data analytics.Read moreRead less
Expanding the Foundation of Planetary Science. Our understanding of the Solar System is based on a foundation of meteorite analyses. Knowing their orbital origin provides a critical spatial context, but we have this data for <0.1% of samples. This project aims to address this issue. There are 66 meteorite falls across Australia with orbits determined by the Desert Fireball Network that await recovery - more than the current global dataset. This project expects to generate new knowledge by applyi ....Expanding the Foundation of Planetary Science. Our understanding of the Solar System is based on a foundation of meteorite analyses. Knowing their orbital origin provides a critical spatial context, but we have this data for <0.1% of samples. This project aims to address this issue. There are 66 meteorite falls across Australia with orbits determined by the Desert Fireball Network that await recovery - more than the current global dataset. This project expects to generate new knowledge by applying an innovative search methodology using drones and machine learning. Expected outcomes include dramatically increasing the number of orbital meteorites. This should provide significant benefits. By linking meteorites to their parent asteroids every rock becomes a small sample-return mission.Read moreRead less
Linkage Infrastructure, Equipment And Facilities - Grant ID: LE0347079
Funder
Australian Research Council
Funding Amount
$160,000.00
Summary
Surface and strain measurement facilities for the investigation of intelligent CAD approaches. The basis of machine learning approaches is the ability to learn or train a system from data gathered through experiments or experience. A major short coming in the development and application of such methods is the lack of good quantitative data. Here we propose the acquisition of dimensional and strain measurement facilities that will allow the investigation of such methods in the context of manufact ....Surface and strain measurement facilities for the investigation of intelligent CAD approaches. The basis of machine learning approaches is the ability to learn or train a system from data gathered through experiments or experience. A major short coming in the development and application of such methods is the lack of good quantitative data. Here we propose the acquisition of dimensional and strain measurement facilities that will allow the investigation of such methods in the context of manufacturing - in particular sheet metal components for the automotive industry. The facilities will enable a database of dimensional and strain information to be established in support of related manufacturing R&D projects.Read moreRead less
Self-supervised feature learning for rapid processing of marine imagery. Fast and reliable quantitative estimates of marine environmental health are needed for scientific studies, design and management of protected areas, and regulatory compliance of industrial activity in the ocean. Australia is collecting seafloor images at increasing rates but expert annotations are not keeping up, meaning that typical machine learning approaches struggle. This project will develop self-supervised techniques ....Self-supervised feature learning for rapid processing of marine imagery. Fast and reliable quantitative estimates of marine environmental health are needed for scientific studies, design and management of protected areas, and regulatory compliance of industrial activity in the ocean. Australia is collecting seafloor images at increasing rates but expert annotations are not keeping up, meaning that typical machine learning approaches struggle. This project will develop self-supervised techniques that use large amounts of unlabeled data to enhance performance. Our design takes advantage of additional information available for marine imagery such as geolocation and remote sensing context. We will explore how these representations can guide additional sampling and improve performance in classification tasks.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE140101749
Funder
Australian Research Council
Funding Amount
$379,480.00
Summary
A computational network model of the mental lexicon. Understanding a word's meaning is a challenge when learning a language and a capacity that is seriously affected in various disorders such as Alzheimer's disease, however little is known about how meaning is organised in the mental lexicon and evolves from childhood into old age. This project aims to build a detailed computational model integrating information available through the senses and structure in the language environment to derive a l ....A computational network model of the mental lexicon. Understanding a word's meaning is a challenge when learning a language and a capacity that is seriously affected in various disorders such as Alzheimer's disease, however little is known about how meaning is organised in the mental lexicon and evolves from childhood into old age. This project aims to build a detailed computational model integrating information available through the senses and structure in the language environment to derive a lexicon that covers most words people know. By distinguishing qualitative different types of meaning relations, this project will allow the prediction of the kind of information and processes required to understand words and an understanding of how this lexicon grows in childhood and declines in old age.Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE240100144
Funder
Australian Research Council
Funding Amount
$444,447.00
Summary
Universal Model Selection Criteria for Scientific Machine Learning. This project aims to develop provably reliable universal model selection criteria to facilitate trustworthy scientific machine learning. Combining stochastic methods with an innovative geometric approach to basic statistical principles, this project expects to characterise, combine, and refine the most successful heuristics for designing and training huge models, such as deep neural networks, into a cohesive theoretical framewor ....Universal Model Selection Criteria for Scientific Machine Learning. This project aims to develop provably reliable universal model selection criteria to facilitate trustworthy scientific machine learning. Combining stochastic methods with an innovative geometric approach to basic statistical principles, this project expects to characterise, combine, and refine the most successful heuristics for designing and training huge models, such as deep neural networks, into a cohesive theoretical framework. The expected outcomes include a general toolkit for assisting neural network design at the forefront of scientific applications. This should significantly improve the quality of scientific predictions by facilitating confident adoption of deep learning methods into the pantheon of trustworthy modeling techniques. Read moreRead less
Stochastic Construction of Error Correcting Codes with Application to Digital Communications. Modern society would be unrecognisable without error correcting codes; mobile telephones, storage devices such as DVD's and high speed data communications simply would not exist. Yet most theoretical results on error correcting codes are asymptotic in nature and ignore computational complexity issues, that is, they are not representative of many real life situations. By building on recent breakthrough ....Stochastic Construction of Error Correcting Codes with Application to Digital Communications. Modern society would be unrecognisable without error correcting codes; mobile telephones, storage devices such as DVD's and high speed data communications simply would not exist. Yet most theoretical results on error correcting codes are asymptotic in nature and ignore computational complexity issues, that is, they are not representative of many real life situations. By building on recent breakthroughs in statistics and stochastic optimisation, this project will develop algorithms for designing optimised error correcting codes subject to realistic finite data length and computational complexity constraints. Successful outcomes will lead to enhanced data communications and storage, greatly benefiting industry and consumers alike.
Read moreRead less