Learning Semi-Naive Bayesian Classifiers from Numeric Data. This project addresses research priority 3, offering frontier technologies. It will deliver better and faster classification technologies that greatly help accomplish many real-world tasks including medical diagnosis, fraud detection, spam filtering and webpage search, where accurate and fast classification is critical to save life, increase efficiency, reduce crime and conserve resources. Hence this project addresses priority 4 as well ....Learning Semi-Naive Bayesian Classifiers from Numeric Data. This project addresses research priority 3, offering frontier technologies. It will deliver better and faster classification technologies that greatly help accomplish many real-world tasks including medical diagnosis, fraud detection, spam filtering and webpage search, where accurate and fast classification is critical to save life, increase efficiency, reduce crime and conserve resources. Hence this project addresses priority 4 as well, better safeguarding Australia from disease and crime. This project will also support a young research group of international standing. It will train the involved researchers to attain a high level of proficiency and excellence in machine learning research and development.Read moreRead less
Discovering justified knowledge from data. Knowledge discovery from data has assumed a critical role in numerous areas of science, commerce and public administration. However, its effectiveness is limited by the undesirable propensity of current techniques to make many false, as well as real, discoveries. This research will rectify that problem, a critical outcome given the potential cost of making decisions or setting policy using flawed information. For example, it may prevent the adoption of ....Discovering justified knowledge from data. Knowledge discovery from data has assumed a critical role in numerous areas of science, commerce and public administration. However, its effectiveness is limited by the undesirable propensity of current techniques to make many false, as well as real, discoveries. This research will rectify that problem, a critical outcome given the potential cost of making decisions or setting policy using flawed information. For example, it may prevent the adoption of ineffective strategies for addressing land degradation; inappropriately targeted public health expenditure; expensive development and clinical trialing of drugs which prove ineffective; and wasted police and security investigations into unfounded suspicions of criminal or terrorist activity.Read moreRead less
Improving human reasoning with causal Bayes networks: a multimodal approach. This project aims to improve human causal and probabilistic reasoning about complex systems by taking a user-centric, multimodal, interactive approach. The project will explore new integrated visual and verbal ways of explaining a causal probabilistic model and its reasoning, to reduce known human reasoning difficulties, and investigate how to reduce cognitive load by prioritising the most useful user- and context-speci ....Improving human reasoning with causal Bayes networks: a multimodal approach. This project aims to improve human causal and probabilistic reasoning about complex systems by taking a user-centric, multimodal, interactive approach. The project will explore new integrated visual and verbal ways of explaining a causal probabilistic model and its reasoning, to reduce known human reasoning difficulties, and investigate how to reduce cognitive load by prioritising the most useful user- and context-specific information. Expected outcomes include novel AI methods that empower users to drive the reasoning process and strengthen trust in the system’s reasoning. Performance will be assessed in medical and legal domains, with significant potential benefits to end users from better, more transparent reasoning and decision making.Read moreRead less
Explaining the outcomes of complex computational models. This project aims to develop new algorithms that automatically generate explanations for the results produced by complex computational models. In recent times, these models have become increasingly accurate, and hence pervasive. However, the reasoning of Deep Neural Networks and Bayesian Networks, and of complex Regression models and Decision Trees is often unclear, impairing effective decision making by practitioners who use the results o ....Explaining the outcomes of complex computational models. This project aims to develop new algorithms that automatically generate explanations for the results produced by complex computational models. In recent times, these models have become increasingly accurate, and hence pervasive. However, the reasoning of Deep Neural Networks and Bayesian Networks, and of complex Regression models and Decision Trees is often unclear, impairing effective decision making by practitioners who use the results of these models or investigate the decisions made by the systems. Practical benefits of clear decision making reasoning by complex computational models include reduced risk, increased productivity and revenue, appropriate adoption of technologies including improved education for practitioners, and improved outcomes for end users. Significant benefits will be demonstrated through the evaluations with practitioners in the areas of healthcare and energy.Read moreRead less
Fault detection and identification in nonlinear complex systems. Complex systems usually comprise a large number of inter-dependent subsystems linked together to perform a certain task. Examples of such systems are power systems, irrigation systems, air traffic control systems, to name a few. Such systems are subject to component failure or malfunction. Total failure can cause an unacceptable financial losses and/or danger to personnel. It is therefore extremely essential, from economic and safe ....Fault detection and identification in nonlinear complex systems. Complex systems usually comprise a large number of inter-dependent subsystems linked together to perform a certain task. Examples of such systems are power systems, irrigation systems, air traffic control systems, to name a few. Such systems are subject to component failure or malfunction. Total failure can cause an unacceptable financial losses and/or danger to personnel. It is therefore extremely essential, from economic and safety view points, that a way be found to ensure reliable and viable operation of complex plants. A first step in achieving this goal is to detect faults on-line and in real-time when they occur and identify their location and characteristics, which is the aim of this project.Read moreRead less
Advanced Bayesian Networks for Epidemiology. We will demonstrate the potential of advanced Artificial Intelligence for medical informatics by extending the capabilities of Bayesian Networks. Bayesian Networks excel when researchers need to combine causal and diagnostic reasoning in areas characterised by uncertainty. But they have one flaw which hinders their use: they do not yet easily mix continuous and discrete variables. We will extend them to handle such mixes, then demonstrate how much the ....Advanced Bayesian Networks for Epidemiology. We will demonstrate the potential of advanced Artificial Intelligence for medical informatics by extending the capabilities of Bayesian Networks. Bayesian Networks excel when researchers need to combine causal and diagnostic reasoning in areas characterised by uncertainty. But they have one flaw which hinders their use: they do not yet easily mix continuous and discrete variables. We will extend them to handle such mixes, then demonstrate how much they can improve on current methods for predicting, among other things, coronary heart disease (CHD).Read moreRead less
Continuous intent tracking for virtual assistance using big contextual data. Recently launched Virtual Assistant products such as Amazon Echo and Google Home are commanded by voice and can call apps to do simple tasks like setting timers and playing music. The next-generation virtual assistants will recommend things to be done proactively rather than waiting for commands passively. This project aims to develop algorithms that can predict what a user intends to do and therefore help virtual assis ....Continuous intent tracking for virtual assistance using big contextual data. Recently launched Virtual Assistant products such as Amazon Echo and Google Home are commanded by voice and can call apps to do simple tasks like setting timers and playing music. The next-generation virtual assistants will recommend things to be done proactively rather than waiting for commands passively. This project aims to develop algorithms that can predict what a user intends to do and therefore help virtual assistants make recommendations that suit users’ needs accurately. It will benefit many service industry sectors of Australia by enabling virtual assistants to provide services proactively.Read moreRead less
On Effectively Answering Why and Why-not Questions in Databases. While the performance and functionality of database systems have gained dramatic improvement, research on improving usability still remains far behind, which results in huge cost of technical support to organisations. This project aims to improve the usability of database systems by effectively answering users' why and why-not questions on query results. This project will invent a novel and generalised model for expressing both the ....On Effectively Answering Why and Why-not Questions in Databases. While the performance and functionality of database systems have gained dramatic improvement, research on improving usability still remains far behind, which results in huge cost of technical support to organisations. This project aims to improve the usability of database systems by effectively answering users' why and why-not questions on query results. This project will invent a novel and generalised model for expressing both the why and why-not questions, efficient strategies for answering questions for complex queries and databases, and novel solutions to scenarios that involve multiple queries. The project will contribute greatly to the fundamental research in query refinement and deliver significant impact on related technology development. Read moreRead less
Multi-Ontologies meet UML: Improving the Software Engineering of Multi-Agent Systems. Multi-agent systems are a new style of software well suited for open, dynamic, distributed, global, heterogeneous environments such as the Internet. Systematic methods are needed to allow multi-agent systems to reason effectively with high level knowledge. This research draws on software engineering practice to develop a theory and methodology for multi-ontologies for expressing knowledge within multi-agent sys ....Multi-Ontologies meet UML: Improving the Software Engineering of Multi-Agent Systems. Multi-agent systems are a new style of software well suited for open, dynamic, distributed, global, heterogeneous environments such as the Internet. Systematic methods are needed to allow multi-agent systems to reason effectively with high level knowledge. This research draws on software engineering practice to develop a theory and methodology for multi-ontologies for expressing knowledge within multi-agent systems that facilitate adaptation and change.
Read moreRead less
Extending association rule discovery to numeric data. This project tackles a key limitation of association-rule discovery, which is one of the main techniques used in data mining. Much valuable data is numeric. However, association-rule discovery cannot satisfactorily model numeric data, a limitation that has greatly restricted its application. This project investigates a novel new technique that overcomes this limitation. Impact-rule discovery finds associations with numeric distributions. ....Extending association rule discovery to numeric data. This project tackles a key limitation of association-rule discovery, which is one of the main techniques used in data mining. Much valuable data is numeric. However, association-rule discovery cannot satisfactorily model numeric data, a limitation that has greatly restricted its application. This project investigates a novel new technique that overcomes this limitation. Impact-rule discovery finds associations with numeric distributions. This allows data analysts to discover precisely the type of information that they usually seek from numeric data, for example, how to maximize either average or aggregate measures of outcomes such as health, compliance, profit, or accuracy.Read moreRead less