Learning to Reason in Reinforcement Learning. Deep Reinforcement Learning (RL) uses deep neural networks to represent and learn optimal decision-making policies for intelligent agents in complex environments. However, most RL approaches require millions of episodes to converge to good policies, making it difficult for RL to be applied in real-world scenarios taking significant resources. This project aims to equip RL with capabilities such as counterfactual reasoning and outcome anticipation to ....Learning to Reason in Reinforcement Learning. Deep Reinforcement Learning (RL) uses deep neural networks to represent and learn optimal decision-making policies for intelligent agents in complex environments. However, most RL approaches require millions of episodes to converge to good policies, making it difficult for RL to be applied in real-world scenarios taking significant resources. This project aims to equip RL with capabilities such as counterfactual reasoning and outcome anticipation to significantly reduce the number of interactions required, improve generalisation, and provide the agent with the capability to consider the cause-effects. These improvements would narrow the gap between AI and human capabilities and broaden the adoption of RL in real-world applications.Read moreRead less
Towards knowledge discovery from imperfect and evolving data. Information extraction from data is critical, both to analyse and protect consumer data. However, many learning techniques are developed using perfect, static datasets, quite different to messy, ever-changing real-world data. This project aims to develop data analytics techniques that can extract accurate information in complex structures from imperfect/incomplete data that changes over time. Expected outcomes are a prototype tool, te ....Towards knowledge discovery from imperfect and evolving data. Information extraction from data is critical, both to analyse and protect consumer data. However, many learning techniques are developed using perfect, static datasets, quite different to messy, ever-changing real-world data. This project aims to develop data analytics techniques that can extract accurate information in complex structures from imperfect/incomplete data that changes over time. Expected outcomes are a prototype tool, tested on real datasets, that combines new techniques in data modelling, algorithm development, and system design. Likely benefits are enhanced Australia's competence in data science through student training and new, robust data tools relevant to critical sectors such as cybersecurity, healthcare, and defence.Read moreRead less
Small Scalable Natural Language Models using Explicit Memory. Deep neural networks have had spectacular success in natural language processing, seeing wide-spread deployment as part of automatic assistant devices in homes and cars, and across many valuable industries including finance, medicine and law. Fueling this success is the use of ever larger models, with exponentially increasing training resources, accompanying hardware and energy demands. This project aims to develop more compact models ....Small Scalable Natural Language Models using Explicit Memory. Deep neural networks have had spectacular success in natural language processing, seeing wide-spread deployment as part of automatic assistant devices in homes and cars, and across many valuable industries including finance, medicine and law. Fueling this success is the use of ever larger models, with exponentially increasing training resources, accompanying hardware and energy demands. This project aims to develop more compact models, based on the incorporation of an explicit searchable memory, which will dramatically reduce model size, hardware requirements and energy usage. This will make modern natural language processing more accessible, while also providing greater flexibility, allowing for more adaptable and portable technologies.Read moreRead less
Short Sequence Representation Learning with Limited Supervision . Predicting events based on short text and video data is widely found in real-world applications such as online crime detection, cyber-attack identification, and public security protection. However, to develop such an effective prediction model is very difficult due to the problems such as limited supervision, heterogeneous multiple sources, and missing and low-quality data. This project is to tackle these challenges. Expected outc ....Short Sequence Representation Learning with Limited Supervision . Predicting events based on short text and video data is widely found in real-world applications such as online crime detection, cyber-attack identification, and public security protection. However, to develop such an effective prediction model is very difficult due to the problems such as limited supervision, heterogeneous multiple sources, and missing and low-quality data. This project is to tackle these challenges. Expected outcome of this project will lay a theoretical foundation for effective short sequence representation learning and build next-generation intelligent systems. This should benefit our society and economy through the applications of multimodality-integrated video technologies for cybersecurity and public safety. Read moreRead less
Discovery Early Career Researcher Award - Grant ID: DE140100772
Funder
Australian Research Council
Funding Amount
$393,414.00
Summary
Response Time Constraints on Category Learning. Theories of associative learning and decision-making are among the most mathematically well developed in psychology. However, theories of learning do not account for the time course of decision-making, and theories of decision-making do not account for how decision-relevant information is learned. This project will develop an integrated theoretical framework linking core principles of associative learning theories with sequential sampling models of ....Response Time Constraints on Category Learning. Theories of associative learning and decision-making are among the most mathematically well developed in psychology. However, theories of learning do not account for the time course of decision-making, and theories of decision-making do not account for how decision-relevant information is learned. This project will develop an integrated theoretical framework linking core principles of associative learning theories with sequential sampling models of the time course of decision-making. The new theory will provide a quantitative account of how incremental associative learning processes drive changes in cognitive representations that, in turn, account for known changes in the time course of decision-making.Read moreRead less
Computer-assisted Clinical Guidelines For The Management Of Manifestations Of Anxiety, Aggression And Depression
Funder
National Health and Medical Research Council
Funding Amount
$354,032.00
Summary
This project focuses on creating a new approach to integration of clinical guidelines and the development of a computer-assisted tool to support medical reasoning in psychogeriatrics. The primary focus of this research is on helping medical practitioners to better manage dementia patients with symptoms of anxiety, aggression and depression living in nursing homes. It has the potential to fundamentally improve the way guidelines are utilised in clinical practice
Discovery Early Career Researcher Award - Grant ID: DE200101253
Funder
Australian Research Council
Funding Amount
$349,586.00
Summary
Making Machine Learning Fair(er). This project aims to develop and implement statistical methods to fight against algorithm bias. In doing so, this project expects to generate new knowledge in the mathematical sciences by employing innovative and interdisciplinary approaches to the development of fairness constraints on machine learning algorithms. Fairness will be seen through the lens of invariance, allowing the developed conceptual framework to find broad applications. Expected outcomes of t ....Making Machine Learning Fair(er). This project aims to develop and implement statistical methods to fight against algorithm bias. In doing so, this project expects to generate new knowledge in the mathematical sciences by employing innovative and interdisciplinary approaches to the development of fairness constraints on machine learning algorithms. Fairness will be seen through the lens of invariance, allowing the developed conceptual framework to find broad applications. Expected outcomes of this project include improved techniques for imposing invariance on deep learning algorithms. This should provide significant benefits to the general public by contributing to the advancement of socially responsible and conscientious machine learning.Read moreRead less
Discrimination learning in humans: Associative and attentional mechanisms. This project offers three major benefits: (1) Australian researchers excel in cognitive neuroscience, learning and psychopharmacology, areas based largely on animal models of human cognition. This project contributes to these areas by specifying the relationship between animal learning and human cognition; (2) the project enhances Australia's international reputation in these areas via its collaboration with a scientist ....Discrimination learning in humans: Associative and attentional mechanisms. This project offers three major benefits: (1) Australian researchers excel in cognitive neuroscience, learning and psychopharmacology, areas based largely on animal models of human cognition. This project contributes to these areas by specifying the relationship between animal learning and human cognition; (2) the project enhances Australia's international reputation in these areas via its collaboration with a scientist of Geoff Hall's stature; it also offers students outstanding research training and international exposure; (3) given Chris Mitchell's industry experience and the relevance of this work to advertising/marketing, this project will generate knowledge relevant to, and possible future collaborations with, Australian industries.Read moreRead less
The Development of Creative Thinking Ability in TAFE Design Students. The aim is to develop an instructional framework as a conceptual basis for instruction aimed at TAFE design student use of imagery in creative thinking. The research addresses the challenges of recent post-compulsory education initiatives, which include a more specific focus on the development of creative problem-solving skills, but provide little guidance on instructional strategies. The research will synthesise the framewo ....The Development of Creative Thinking Ability in TAFE Design Students. The aim is to develop an instructional framework as a conceptual basis for instruction aimed at TAFE design student use of imagery in creative thinking. The research addresses the challenges of recent post-compulsory education initiatives, which include a more specific focus on the development of creative problem-solving skills, but provide little guidance on instructional strategies. The research will synthesise the framework from theories about cognitive processes, creative problem-solving and the use of mental imagery; and the results of an imagery training program and creative problem-solving interventions. The outcome will be an instructional framework for guiding the development of student creative thinking in design-based courses in TAFE.Read moreRead less
Evaluating models of category learning that use general feature-based representations. Three competing models of human category learning will be evaluated by comparing their behaviour to human performance on an experimental task where each model makes qualitatively different predictions. A series of theoretical and algorithmic advances will be undertaken to ensure each of the category learning models uses the same feature-based representation. Because the three models propose very different lear ....Evaluating models of category learning that use general feature-based representations. Three competing models of human category learning will be evaluated by comparing their behaviour to human performance on an experimental task where each model makes qualitatively different predictions. A series of theoretical and algorithmic advances will be undertaken to ensure each of the category learning models uses the same feature-based representation. Because the three models propose very different learning processes, their comparison will give insight into the basic cognitive process of categorisation. The algorithms for generating feature representations and modelling human category learning will also have potential for application in data visualisation and information handling systems.Read moreRead less