<<

, and Data — UQ capabilities

March 2021 Artificial Intelligence, Machine Learning and Data Science

Artificial Intelligence, Machine Learning and Data Science

Contents 4 Research Strenghts

7 Health & Medicine

10 Society & Government

11 Business

13 Mining, Agriculture, Defence & Energy

17 AI Education & Outreach UQ Research Partnerships [email protected] 18 Phone: +61 7 3365 3559 Our People

Page 2 Artificial Intelligence, Machine Learning and Data Science

The University of Queensland recognises the long-term implications of AI and is committed to driving innovations in AI research and its application.

Introduction

Advances in computer power and Artificial Intelligence at UQ big-data analysis or digitalisation AI capability at UQ cuts across both have underpinned the pathway to fundamental and applied research, as Artificial Intelligence (AI), which well as a broad range of outreach and can be described as the capacity or educational activities. ability of a machine to learn and solve problems mimicking the human mind. AI research and its applications can be broadly categorised into research Digital transformation is dramatically strengths, impact areas and services. changing the way business operates, The research strengths of UQ in AI, engaging almost every sector of the include machine learning (ML) and economy along the way. data science (DS) and present a strong Artificial Intelligence is also a powerful portfolio of fundamental knowledge tool to improve business processes critical for the future of AI. and will have a global impact on At the same time, there is a significant the future workforce through the capability at UQ on key impact areas transformation of the relationship that target solutions for complex between people and technology. global problems across multiple Advancements in AI are expected to industry sectors. contribute as much as US$15.7 trillion to the global economy by 2030 and AU$315 billion to the Australian economy by 2028. The University of Queensland recognises the long-term implications of AI and has invested strongly in building capability in AI research and its application.

Page 3 Artificial Intelligence, Machine Learning and Data Science

Research strengths

Detecting fake news Data Management and Data Information resilience Science Multiple industry reports and recent through AI media coverage identified ‘data gone The ability to algorithmically Big-data analysis wrong’ as the biggest risk factor for AI spread false information Big-data analysis relates to and other , and through online social networks conducting multi disciplinary its impact is increasingly recognised together with the data-driven research on large-scale, as a threat from mundane cyber- ability to profile and micro- unstructured, multi-modal and criminals to sophisticated well-funded target individual users has made complex data to find innovative and entities, raising concerns of national it possible to create customised practical solutions, and to create security. Information resilience is false content. value from big data in business, the capacity of organisations to Fortunately, similar data-driven scientific and social applications. build, protect, and sustain agile data and AI methods can also be Dr Hongzhi Yin is the current pipelines, capable of detecting and used to detect misinformation director of the Responsible and responding to failures and risks and to control its spread. Sustainable Big Data Intelligence across the value chain in which the Dr Gianluca Demartini has (RSBDI) Lab. The RSBDI lab strives data is sourced, shared, transformed, partnered with Facebook to: to develop energy-efficient, privacy- analysed, and consumed. The success of AI’s implementation will require • develop automatic fake preserving, robust, explainable robust mechanisms and capacity news detection systems and fair data mining and machine learning techniques with theoretical building around information resilience. • investigate the possibility backbones to better discover Prof. Shazia Sadiq leads research to of generating information actionable patterns and intelligence address these issues, including: credibility literacy tasks from large-scale, heterogeneous, • Responsible use of data assets that can give people the networked, dynamic and sparse (principled approaches to data skills to recognise fake data. RSBDI lab joins forces with governance, access and sharing) news. urban transportation, healthcare, • Data curation at scale (machine agriculture, e-commerce and Low energy mobile learning, crowd-sourcing and marketing fields to help solve human-in-the-loop techniques) devices the societal, environmental and • Algorithmic transparency economic challenges humanity Browsing the internet involves (approaches to promote faces. transferring user behavioural interpretability, uncertainty data to a large-cloud based Large-scale spatiotemporal data quantification unbiasedness, system, which in turn are Prof. Xiaofang Zhou leads research transparency and reproducibility massive consumers of energy. on large-scale spatiotemporal in the design of learning Dr Hongzhi Yin has developed data processing. He focuses on algorithms) a cost-effective, low-energy developing advance knowledge • Trusted data partnerships (data mobile system to process the and techniques from data quality literacy, trust in data linking, lifting information. management, data storage and barriers to data sharing) This eliminates the need for high-performing processing to data • Creating value from data (data transferring information to mining and ML for trends prediction monetisation, business process in spatial information systems, cloud systems (reducing CO2 improvement, measuring emissions), and protects a user’s intelligent transport systems, IoT, analytics value and organisational private information. and streaming sensory data. structures).

Page 4 Artificial Intelligence, Machine Learning and Data Science

Research strengths

Computer vision and Big Media Data Analytics Information retrieval multimedia computation The Big Media Data Analysis group conducts large-scale multi disciplinary Prof. Guido Zuccon investigates research on unstructured, multi-modal how people are using search and complex data, achieving human- Computer vision concerns itself with engines to seek health advice on understandable machine intelligence the understanding of the real world the web. Widely disseminated by for multimedia, social media, health through the analysis of images. More the media, the work shows that less data analytics and knowledge specifically, it infers properties of the than two-thirds of those who search discovery. online to make health decisions observed world from an image or a The group is developing AI-powered then go on to consult a health collection of images. Computer vision technologies to enable tangible professional with regard to their has attracted considerable attention benefits for the whole-of-society, health conditions. from industry and academia for its wide range of applications, many of ranging from applications such as which have economic value. autonomous vehicles, image/video content understanding and retrieval, Prof. Anders Eriksson and his team event detection to user behaviour lead research on: analysis, and recommender systems. • The Big Media Data Analysis Group • has been engaged in several cross- • disciplinary projects with social science, chemical , and • Geometry processing advanced water management teams • Medical to extend the boundaries of their • Skin cancer diagnosis research and generate impact to other scientific disciplines. Object recognition Object recognition is the ability to recognise what objects an image shows. The technology behind it aims to extract robust features from objects to improve detection and recognition, and provide solutions for real-time applications such as surveillance. Being able to detect and recognise human faces and their emotions from facial expressions is an important part of the recognition process.

Page 5 Artificial Intelligence, Machine Learning and Data Science

Research strengths

Machine learning and d a t a m i n i n g Explainability in Machine

Learning Explainability in ML ML and data mining A/Prof. Helen Huang and Prof. Dr Sen Wang’s research focuses Dr Mahsa Baktashmotlagh’s Xiaofang Zhou lead the Data on developing interpretable deep research focuses on developing Science Group. Using state-of-the- models in many data-rich fields machine learning and data mining art and deep learning techniques, to provide insights into the data, techniques and applying them to they are developing interpretable variables and decision points used to diverse contexts including: deep models in many data-rich make a recommendation. In health- • Visual data analysis (visual fields. related projects, Dr Wang’s research domain adaptation, video classification, and animal’s In health-related projects, they has significant applications in health foraging behavioural seek to achieve high performance informatics and social media. analysis) in medical prediction tasks such as illness severity prediction in ICUs, Theoretical aspects of semi- • Road traffic networks but addressing explainability issues supervised learning (mining large scale road raised by black-box models. Prof. Geoffrey McLachlan is a UQ traffic networks and building expert in semi-supervised learning a road-load balancing tool (SSL), an AI method that analyses to predict congestion on any datasets comprised of both labelled road in the city) and unlabelled training observations. • Biomedical data (prediction SSL aims to classify the unlabelled of neonatal sepsis) data using the existing information of • Finance (hedging foreign the labelled data. exchange trading risks).

Page 6 Artificial Intelligence, Machine Learning and Data Science

Research strengths

O p t i m i s a t i o n a n d s t a t i s t i c s algorithmic approaches will be applied to sustainable fishery management — an important problem for Australia Novel optimisation algorithms and an ideal context for POMDPs. Dr Fred Roosta-Khorasani and The project will advance research his research team design, analyse in artificial intelligence, dynamical and implement novel optimisation systems, and fishery operations, and algorithms for training modern benefit the national economy. machine learning problems. The underlying approaches draw from Evolutionary algorithms and randomised techniques, statistics, optimisation non-convex non-linear analysis, Dr Marcus Gallagher’s research geometry, and parallel/distributed interests include AI and nature- computing. Dr Roosta-Khorasani inspired problem solving techniques is also interested in using novel ML and algorithms. More specifically, models for scientific discovery, in his work focusses on optimisation, particular the design of innovative metaheuristics and evolutionary physics-aware deep learning models computation, machine learning and the development of their and exploratory data analysis and statistical learning theory. visualisation. He uses statistical and probabilistic models and techniques to address these challenges. Sequential decision making under uncertainty Sequential decision making under Statistics and probability uncertainty is a recurrent challenge Dr Yoni Nazarathy’s research expertise in several related fields, including AI, lies in applied mathematics and control theory and statistics. statistics with a focus on stochastic To better understand randomness, operations research, scheduling, Dr Nan Ye and Prof. Dirk Kroese use resource allocation, optimisation, data Monte Carlo simulations. science and computer-controlled systems. Their research aims to develop theoretical and algorithmic tools to address the challenge of scaling-up existing methods to large complex environments, and explore their application in sustainable fishery management. To approach these challenges, they use partially observable Markov decision processes (POMDPs) and Monte Carlo techniques. POMDPs provide a general mathematical framework for sequential decision making under uncertainty. Both theoretical and

Page 7 Artificial Intelligence, Machine Learning and Data Science

Research strengths

Human centred AI and Emotions and e-wearables Language technologies design The detection of emotions using smartphones/wearable sensors and Language technologies studies Extended reality (XR) personal informatics is becoming how computer programs can increasable popular. Dr Chelsea XR is a technology that covers both analyse, produce, and respond Dobbins’s research focuses on the Mixed Reality (AR) and Virtual Reality to human texts and speech, area of lifelogging, i.e. how our lives (VR), Dr Arindam Dey’s research and develop tools to enable can be logged through the use of aims to design novel interfaces using computers to recognise human wearable sensors and mobile devices. both AR and VR, individually and language. combined (XR) in multiple application In this context, she uses AI methods to: Prof. Janet Wiles leads domains including but not limited • Visualise unconscious emotional research in technologies to remote collaborations, effective processes and increase self- involving language recognition, computing, gaming, education, health, awareness to promote positive processing, production and and sustainability in collaborative coping strategies to protect future evaluation. applications. health. The Centre of Excellence for • Detect emotions using the Dynamics of Language physiological and neurological Visual SenseMaking (CoEDL) is an umbrella group for signals to make virtual and language projects. Dr Mashhuda Glencross’s research augmented reality interfaces The centre is involved with group aims to create visual adaptive and communicative of several language technologies mechanisms to understand large data such emotional states. such as Elpis speech-to-text, sets from multiple sources, develop meaningful visuals, narratives and build Opie social robots, and Florence Bias and Fairness (supporting communication for literacy about the data. Her research AI sometimes makes systematic people living with dementia). team focuses on visualisation, energy data, sense making, human-in-the- errors that may lead to bias and unfair loop, virtual reality, simulation, visual decisions. At UQ, we investigate analysis, and storytelling. They use AI human-centred AI methods to better methods to narrow the scope of what understand the causes of and to data to visualise and to aid automatic better deal with bias and fairness in creation of techniques and metaphors AI. Examples of our research include: to experience the data. studies on online misinformation aimed at understanding how it spreads and how it is perceived by people, AI models to understand and predict online user interaction behaviour, and methods to reduce AI mistakes that are difficult to be automatically detected.

Page 8 Artificial Intelligence, Machine Learning and Data Science

Health & Medicine Impact areas

Digital pathology Healthcare and medicine continues to be an

UQ’s high-performance outstanding impact area for UQ with significant computer, Weiner, is now able background IP and world-class research leadership. to deliver unprecedented performance for image Digital pathology Neuroengineering and rehabilitation processing and deep Prof. Brian Lovell leads UQ’s Digital Dr Alejandro Melendez-Calderon’s learning algorithms, allowing Pathology team. The team aims to research team aims to understand advancements in digital develop technologies that make and aid the recovery of, assistance pathology. pathology workflows faster and more or replacement of, functions The team behind this initiative efficient. lost because of neurological includes UQ’s Professor diseases or injuiries such as stroke, Prof. Lovell has collaborated with Lovell and Dr Arnold Wiliem, traumatic brain injury, Parkinson’s Sullivan Nicolaides to continuously as well as Peter Hobson disease, or spinal cord injury. create computer image datasets from and Anthony Jennings from This is an interdisciplinary field different pathology areas to further Sullivan Nicolaides Pathology. of research at the interface of advance the field. neuroscience, engineering and AI, Medical imaging and deals with the understanding of neurological principles and Prof. Stuart Crozier’s expertise lies in design of technology that is imaging technology and applications meant to improve the quality of instrumentation for physiological life of people with disabilities. measurement and semi-automated

diagnostics. Brain development and information The commercial and academic processing impact of his work in Magnetic Resonance Imaging (MRI) has been Prof. Geoffrey Goodhill is interested significant. in how brains process information, particularly during development. About two-thirds of all high-end, clinical MRI systems installed This includes how growing nerve worldwide after 1997 contain fibres use molecular cues to make patented technology co-invented guidance decisions, how map- and developed by Prof. Crozier. like representations of visual inputs form in the optic tectum Movement neuroscience and and visual cortex, and how these computational neuromechanics maps code sensory information. Prof. Paul Hodges’ research team aims His research team uses a combination to understanding neuromechanical of mathematical, experimental mechanisms that underlie healthy and computational techniques to and impaired sensorimotor function. help test hypotheses about the Our approach involves the use of biological mechanisms that drive machine learning and AI techniques these changes during development. to generate computational models of human function and behaviour.

Page 9 Artificial Intelligence, Machine Learning and Data Science

Health & Medicine Impact areas

Laser technology and early collecting hospital data from 64 detection of cancer countries around the world. In Brain activity and Prof. Aleksandar Rakić leads UQ’s collaboration with Oxford University, functional MRI and Microwave Engineering the UQ data team led by Dr Sally group. Shrapnel aims to: Prof. Markus Barth is • Characterise acute kidney developing very fast functional His research focuses on the injury in COVID-19 patients. MRI techniques with the development of technologies highest spatial resolution for sensing and imaging across • Develop a predictive algorithm possible. the electromagnetic spectrum to identify COVID-19 patients at including microwave, terahertz high risk of kidney injury He is particularly interested in wave and optical systems. identifying the small functional • Validate and deploy this His research team is developing units of the brain, such as algorithm in a resource poor disrupting technologies based on the cortical layers and columns, setting (Latin America). terahertz quantum cascade lasers and to better understand brain (QCLs) for the next generation of Clinical informatics function. terahertz imaging systems suitable Dr Clair Sullivan is a leading for early detection of skin cancer. Australian researcher on electronic medical records and the digital Predicting from complex data transformation of health systems. Dr Dr Alina Bialkowski is a computer Sullivan leads the Queensland Digital vision, ML and signal processing Health Research Network, which is a algorithms researcher, working on transdisciplinary group of researchers applications in medical imaging and focussed on delivering better intelligent transport systems. healthcare outcomes using digital Her research interests include technology. She specialises in the extracting information from creation and integration of decision complex data to model patterns and support (including AI) into routine developing interpretable models to clinical care. She is an expert in the solve real-world problems. deployment, governance and safety of digital solutions in healthcare. ML for acute kidney injury in Biomedical Applications and Devices COVID-19 patients Dr Philip Terrill leads research on novel COVID-19 has no established medical diagnostic tools and therapies protocols for treatment to kidney with the goal to improve the health injury. Around one third of patients outcomes of people in Australia and hospitalised with COVID-19 have globally and bridge the gap between significant injury to their kidneys. clinical physiology and biomedical We currently have very limited engineering. understanding of how, why, and when the kidneys are injured during His research focusses on the infection with COVID-19, and no application of AI, ML-based solutions proven strategies to prevent the and robotics to the data-driven analysis condition. of medical conditions such as sleep apnoea and osteoarthritis, as well The International Severe Acute as the study of human sensorimotor Respiratory and Infection Consortium control, neuro-rehabilitation and human (ISARIC) - Acute Kidney Injury augmentation technologies. Analysis is an observational study

Page 10 Artificial Intelligence, Machine Learning and Data Science

Health & Medicine Impact areas

Automatic medical image analysis There is a need to use scientific Dr Shakes Chandra leads research on expertise to distinguish patterns automatic medical image analysis. in extremely high-dimensional His expertise includes image feature spaces that are biologically processing, ML, discrete tomography meaningful from those that are and medical image analysis. artefacts of the data generation More specifically, he has developed process, ML, discrete tomography and deformable models and ML medical image analysis. algorithms for knee, hip and shoulder- The Boden group is actively joint segmentation for deployment collaborating with scientists in on the Siemens Syngo platform. genomics, epigenetics and protein UQ is applying the automatic medical science, and is regularly developing image analysis tool to: methods, embedding ML and AI • Complete 3D reconstruction algorithms, that are then used by the of individual bodies with every scientific community. mole marked, analysed, and compared with previous scans to assess and diagnose any AI in skin cancer diagnostics potential skin cancer Australia has the highest rate of skin • Develop AI and ML-based cancer in the world, with hundreds of algorithms to automate the thousands of new diagnosed cases analyses of clinical imaging and each year, and early detection is to model the biomechanics one of the most important factors of joint function, while in their prognosis. Today the most translating such technologies effective method for improving the for industry partners such as early detection rate is patient skin Siemens Healthcare, Germany. self-examination complemented by physician directed examination. Prof. Bioinformatics Anders Eriksson and his team are developing technology to automate Bioinformatics draws on computer patient skin self-examination, science, math and statistics to enable providing the means to conduct discoveries in molecular biology data. convenient and personal full-body Of particular interest is the application scans for the purpose of ongoing of AI and machine learning (ML) that AI driven monitoring of skin health. promises to leverage the information Presenting the general public with this hidden in massive datasets that are ability would allow skin cancer to be currently being generated by genome detected at a much earlier stage and sequencing based technologies. thereby dramatically increasing overall A/Prof. Mikael Boden and his group’s patient survival rates. aims to investigate, develop and apply the theory and practice of AI, statistical ML, data mining and probabilistic methods to understand and resolve a range of open problems in genomics, molecular and systems biology.

Page 11 Artificial Intelligence, Machine Learning and Data Science

Health & Medicine

Wrist-worn devices including clinical research and consumer devices can provide a high volume of meaningful data about everyday routines.

Machine learning and ML and decision-making Dr Tennakoon Mudiyanselage’s d a t a m i n i n g research uses health and administrative data to develop AI/ML and accelerometry devices explainable ML learning models Prof. Simon Smith leads the to support decision-making. She Sleep and Health Group, with has applied unsupervised ML the broad aim to bridge the methods to develop effective biological and social approaches to discover knowledge for community benefit. from social media networks based His research team uses AI/ML on user interaction patterns. to analyse accelerometry and In one of her projects at the Institute other indicators of behavioural for Social Sciences Research (ISSR), rhythmicity in natural settings she explored the effectiveness of to differentiate and understand these methods to identify children, sleep-wake and activity states. admitted to paediatric intensive care Wrist-worn devices including clinical units in Queensland, at risk of poor research and consumer devices can future educational outcomes. provide a high volume of meaningful data about everyday routines. Prof. Smith’s research team leads a study seeking to classify high- frequency activities (e.g. on-road driving) from actigraphy using contemporary AI/ML approaches.

Page 12 Artificial Intelligence, Machine Learning and Data Science

Society & Government Impact areas

Machine vision and social responsibility Critical parallel and complementary areas Social platforms currently demonstrate UQ’s capacity in end-to-end solutions include ‘machine vision’ systems that automatically classify and recognition of societal/business challenges in AI. faces, expressions, objects, and brand logos in images. Data from machine vision systems Data for action in government Uber drivers, e-government and is used to provide targeted Finding ways that machine learning addressing inequalities associated content to users, often without tools can support high stakes with through basic income their knowledge and sufficient decisions by government agencies is proposals. He is also interested in public oversight. Dr Nicholas the focus of research by Prof. Rhema public policy approaches to regulating Carah uses a novel combination Vaithianathan and her team at the AI to reduce bias and increase of computational and cultural transparency in algorithmic decision- research methods to: Centre for Social Data Analytics. making. • Examine how machine Prof. Vaithianathan works closely vision works in platforms with agencies to develop and implement machine learning tools such as Instagram AI in the public discourse that harness existing data to support • Explore their role in Dr Stan Karanasios’s research focuses everyday visual contexts decision making in health and human services. The use cases for these on how organisations use, adapt, and (festivals, food, and lifestyle navigate new digital technologies. sports) tools are helping agencies to address the challenges of child maltreatment, He is currently working with a leading • Improve public homelessness and elder abuse. European car manufacturer to understanding of machine examine ways of re-imagining work vision systems. From the early stages of these translational research projects, Prof. process to accommodate increasingly Vaithianathan’s team focus is on the digital products. Protecting children effort on the human concerns that Prof. Rhema Vaithianathan determine community comfort (like AI and government accountability and her team developed the equity, ethics and transparency of world-first Allegheny Family the tools) as well as, the data science Governments have been using digital Screening tool (AFST), a child itself. technologies for decades, with maltreatment decision support implications for the operation and tool implemented by Allegheny power of the state, and its relationship Social impact of AI and future of County, PA (United States) in with the governance of citizens. work 2016, with the team’s approach Increasingly, algorithmic government widely praised as transparent Prof. Greg Marston’s research interests and AI is heightening concerns with and inclusive. The AFST is a are on the social impact of AI and government accountability and predictive risk modelling tool automation in the future of work. transparency, and the reproduction of that rapidly integrates and He focuses on the welfare-work bias, discrimination and inequalities. analyses hundreds of data nexus and the changing nature Charting these dynamics and elements for each person of employment, including the gig developing policy and governance involved in an allegation of child economy. frameworks is important. maltreatment. He has written various submissions Prof. Paul Henman’s current research and given evidence at government covers the nexus between social policy, inquiries into the future of work. administration and digital information He has undertaken research on technologies.

Page 13 Artificial Intelligence, Machine Learning and Data Science

Society & Government Impact areas

Use of AI by government These resources are commonly organisations created by domain experts. RiPPLE Dr Ida Someh’s research focuses on is a UQ developed and supported Trust and ethics in AI organisational and societal impact AES that takes the crowdsourcing In recent years, Australia has seen of data, analytics and artificial approach of partnering with students, national inquiries into institutional intelligence. also referred to as learner sourcing failures and trust breaches by the to create the resource repository. , To She is currently investigating how banking and financial services date, RiPPLE has been implemented government organisations can develop sector, aged care, churches and in over 70 courses across a range and deploy successful AI applications sporting organisations. Long-term of disciplines including Medicine, with significant public value. sustainability of business depends Pharmacy, Psychology, Education, on trust and goodwill, ethical and Despite its potential, AI uptake is still Business, and trustworthy organisational conduct low. The study analyses the challenges Biosciences. requires robust governance of AI uptake, and how to mitigate systems that benchmark and them. Intelligent learning analytics assess performance and culture. The process involves qualitative dashboards Prof Nicole Gillespie co-leads surveys to industry contacts that have the Trust, Ethics and Governance delivered AI solutions to the public Learning analytics dashboards Alliance (TEGA) initiative aiming sector. commonly visualise data about to address complex and rapidly students with the aim of assisting After gaining insight on the types of evolving challenges in society, such students and educators in AI projects carried out and the most as the rise of advanced technology understanding and making informed common challenges experienced, a and artificial intelligence, and decisions about the learning process. deeper analysis into a select number unprecedented challenges on To assist with making sense of of cases follows through interviews managing trust, ethics and complex and multi-dimensional data, and other in-depth data-collection governance issues. many learning analytics systems and methods such as ML, government AI dashboards have relied significantly implementation, AI explainability and on AI algorithms based on predictive human-AI teaming methods. AI and its application analytics. While predictive models in Education have been successful in many domains, There is a growing consensus that there is an increasing realisation of Educational crowdsourcing to applications of AI in education have the inadequacies of using predictive support personalisation a transformational impact on the models in decision-making tasks that Educational crowdsourcing to educational landscape. The UQ AI affect individuals without human support personalisation: Adaptive in Education research group led oversight. To address this challenge, educational systems (AESs) make by Dr Hassan Khosravi draws on Dr Khosravi’s team have developed use of data about students, learning insights from the learning sciences a learning analytics dashboard called processes, and learning products to and techniques from the fields Course Insights that employs a suite of adapt instructions for each student. of human-computer interaction, state-of-the-art algorithms, from online AESs rely on learner models that human-centred AI and learning analytics processing, data mining and capture an abstract representation analytics to design, implement, process mining domains, to present of a student’s ability level based on validate and deliver technological an alternative human-in-the-loop their performance and interactions solutions that contribute to the AI method to enable educators to with the system. AES uses information delivery of learner-centred, data- identify, explore and use appropriate from the learner model to recommend driven learning at scale. interventions for subpopulations of items from a large repository of students with the highest deviation learning resources that best match the in performance or learning process current learning needs of a student. compared to the rest of the class.

Page 14 Artificial Intelligence, Machine Learning and Data Science

Business Impact areas

Implementing complex and even Digital transformation inscrutable Artificial Intelligence Prof. Andrew Burton-Jones’s research Dr Tapani Rinta-Kahilsa’s research group studies how organisations adresses current issues on socio- learn to use new information systems technical change in organisations. effectively. The success of any His focus areas include the organisation digital journey will depend organisational implementation of AI on their use of data and analytics. A and automation, the identification major focus is ensuring their effective and prevention of technologies’ understanding and use of data and negative effects on individuals analytics to improve practices and and organisations, as well as the institutions. He leads a large program of discontinuance of incumbent work with Queensland public hospitals information technologies. on their digital transformation journey. The topics studied range from individual-level studies (how data Digital transactions is input, retrieved/extracted, and its A/Prof. Adrian Athique is an expert in quality maintained), to socio-technical cultural studies in Asia. His research studies (how groups interact with interests include digital media and data, how practices change), and society, transnational media in Asia, through transformation topics (how technology and economy of media organisations and sectors can improve). systems, and visual sociology. He currently leads a project on digital transactions in India.

Page 15 Artificial Intelligence, Machine Learning and Data Science

Mining, Agriculture, Defence & Energy Impact areas

Mining design, development, and evaluation AI and mining AI and mining equipment processes. equipment Prof. Ross McAree’s research group Prof. Robin Burgess-Limerick is has been involved in the development an experienced human factor and The development of AI and ML of several automation technologies ergonomics researcher. techniques has made possible that have been commercialised for the His current projects are in the areas the design of autonomous mining sector. of mining automation-human systems vehicles. These outputs have benefited the integration, as well as equipment Self-driving cars will play an global mining industry by more than design to reduce injury risks, manual important role in the near $1BN. tasks risk management, and whole- function of transportation Prof. McAree led the development and body vibration measurement and systems as they could be more demonstration of the world’s first fully management. safe, efficient and generate autonomous mining excavator as part lower carbon emissions. of a 10-year collaboration with Joy Prof. Ross McAree’s research Global Surface Mining (now part of Agriculture has made contributions to Komatsu). vehicle automation through He has also collaborated with Crop modelling the application of AI/ML/ Caterpillar Inc. to develop and trial the Prof. Graeme Hammer’s research DS techniques to learn about world’s first autonomous bulldozer capabilities focus on the major cereal traction parameters on road capable of production dozing using crops: sorghum, maize and wheat. segments and fault detection the ‘pivot push’ method. Graeme has expertise in crop and identification. ecophysiology and modelling, which Prediction in mining he uses to investigate traits and Prof. Peter Knights’ research relates management systems that have FastStack — to maintenance and reliability the potential to deliver productivity evolutionary computing engineering. He has promoted the gains in water-limited production now-widespread use of logarithmic to stack desirable alleles environments. scatter plots (also known as jack- He leads the UQ link to the APSIM in wheat knife diagrams) to characterise and Initiative, which is responsible for the prioritise downtime events. on-going development of the APSIM FastStack is an AI platform Dr Mohsen Yahyaei leads the modelling software platform, used currently under development Advanced Process Prediction and internationally. (2018–2022) by Prof. Ben Control (APPCo) program, which aims His research team aims to enhance Hayes in collaboration with to transform unit process modelling profitability and sustainability LongReach Plant Breeders and simulation. of cereal and legume cropping to assist the accelerated systems in tropical and sub-tropical production of wheat varieties Human-system integration in environments through the integration with superior performance mining automation of digital agriculture technologies traits. If successful, Australian The introduction of automation to and capabilities at a molecular level, growers will undoubtedly mining has great potential to reduce whole plant, and production systems, benefit as well. safety and health risks by removing with their frontier research in climate people from hazardous situations. modelling, prediction agriculture, For a system to function optimally, genetics and farming systems to human abilities and limitations provide tomorrow’s decision support must be considered in the planning, tools today.

Page 16 Artificial Intelligence, Machine Learning and Data Science

Mining, Agriculture, Defence & Energy Impact areas

ML and crop performance farmers, on farmers’ fields. Innovations in plant testing in His current research areas include: The APSIM initiative Australia (INVITA) is a $AU10M • Adapting to climate and global The Agricultural Production national initiative co-funded by change Systems sIMulator (APSIM) Grains Research and Development is a comprehensive model Corporation (GRDC). INVITA • Food security and poverty developed to simulate leverages off a major EU initiative mitigation across Eastern Africa biophysical processes in project - INVITE (Innovations in • Systems modelling and whole- agricultural systems, particularly plant Variety Testing in Europe) and farm resource allocation. as it relates to the economic partners with Wageningen University and ecological outcomes of and CSIRO. INVITA aims to improve management practices in the the data provided to growers on crop Defence face of climate risk. varieties performance. A team led The APSIM model can explore by Prof Scott Chapman is running Cyber security research and digital options and solutions for the field trials and applying sensing, competitiveness food security, climate change UAVs and IoT sensors together with The digital competitiveness of any adaptation and mitigation machine learning to develop analytics organisation implementing Artificial and carbon trading problem to improve crop performance Intelligence and Big Data analytics domains. APSIM has evolved predictions. depend mostly on high-quality large into a framework containing datasets for training and forecasting, Crop production forecasting systems many of the critical models collected from authoritative and at a regional scale required to explore changes realistic sources. Dr Andries Potgieter is an expert in agricultural landscapes with Advanced cyber-attacks and cyber in the areas of seasonal climate capability ranging from the threats have the power of mutating forecasting, remote and proximal simulation of gene expression and breaking out faster than the sensing with applications in the through to multi-field farms response of current detection models. and beyond. development of crop production outlooks and less risk-prone cropping Prof. Ryan Ko and his research systems across Australia. team use IoT security analytics; Machine learning for industrial control systems security His research focuses on the complex agriculture analytics; network security analytics; integration of remote sensing blockchain security analytics; cloud Dr Anders Eriksson applies ML technologies, spatial production security analytics; cyber security; and deep-learning models to modelling and climate forecasting and vulnerability detection to ensure measure a range of plant traits systems at a regional scale in that research on algorithms and pertinent to crop management agricultural settings. approaches are state-of-the-art. and decision making. Farming systems modelling UQ’s Cyber Security Research Centre He also helps breeders to Prof. Daniel Rodriguez is a leader in and DATA61 are working together to improve plant performance the research and development of create high-quality large cyber security through efficient computer agricultural systems modelling. datasets and explore their innovative vision techniques based on DL and digital-competitiveness use. applied to high-resolution RGB Prof. Rodriguez’s research team has imagery. expertise in the areas of agro-ecology, The study will have two work packages: simulation analysis and modelling, • Establish a massive data agronomy, soil science and root collection environment to physiology. integrate datasets of normal Their research applies systems operations, attack and defence thinking and analysis tools and is scenarios, and APT information conducted in close collaboration with

Page 17 Artificial Intelligence, Machine Learning and Data Science

Mining, Agriculture, Defence & Energy Impact areas

observed across Network and Smart energy Security Operation Centres Power Quality Meter (NOC and SOC) in UQ (and its global partners) and DATA61. Dr Richard Yan aims to apply a Power Quality Meter (PQM) system to • Strive to push the boundary analyse network events. of cyber big data analytics based on the created PQM captures and monitors quality high-quality dataset. data from substations and grid assets. The project will use machine learning (ML) and AI algorithms to extract key Using ML to detect cyber-attacks information considered as events, and Cyber-attacks are becoming verify the event identification method increasingly sophisticated and on the field trial according to an frequent. The estimated average established event library. cost of each successful attack Power utilities do not have the time on business is already estimated and capacity to analyse big data, and at more than a $1M per incident. he is working with them on this topic Dr Marius Portmann has an ongoing to facilitate the process. and successful collaboration with Comscentre to develop an advanced Wide-area power energy system AI based system to detect, identify monitoring and autonomously mitigate cyber- attacks in large-scale computer This research, led by Prof. Tapan Saha, networks. relates to the cyber security of target cyber-physical system (CPS) – WAMS As part of the collaboration, they have (Power system wide-area monitoring developed and implemented an AI system) in electric power grids. platform for the detection of network intrusions and cyber attacks. The focal point of this research is on developing ML data driven-based methods and tools to efficiently Solar panels monitoring and measure (through WAMS) data management authentication at control centres for Dr Rahul Sharma’s research uses AI/ better data reliability and security. ML/DS technology to automatically detect and locate under-performing solar panels in large solar farms. In particular, it uses residential loads to visualise where the spots using more energy are and where there is potential for saving through aggregate load data disaggregation techniques.

Page 18 Artificial Intelligence, Machine Learning and Data Science

Mining, Agriculture, Defence & Energy Impact areas

Transforming big data into ML/AI in market design and problems analysis Making PV system in smart data This research develops and uses ML ML/AI in market designs research use distribution networks techniques to automatically, extract ML methods to construct generative cost-effective and meaningful information from existing models as inputs to simulation studies, greener data and transform it into domain- especially power flow, market analysis, specific knowledge. Outcomes from and design studies to solve a variety The rapid deployment of solar this research can improve power of techniques to solve problems. photovoltaic (PV) systems in system asset condition assessment In a wider context, Dr Archie Chapman distribution networks, while and maintenance. develops and applies principled AI, indicative of the desire to move The research team led by Dr Lakshitha game theory, optimisation and ML away from fossil fuels, has Naranpanawe has created a Bayesian methods to solve large-scale and created new challenges due to information fusion-based algorithm, dynamic allocation, scheduling and the bi-directional power flow which can integrate data from field queuing problems, including: and power fluctuations of PV. condition monitoring, maintenance • Approximate dynamic Professor Tapan’s research records, and failure statistics to programming, reinforcement team aims to provide a more construct an inference probabilistic learning and policy function cost-effective way to monitor model to determine a health index for approximation used to emulate and predict the operating power system assets. decision-makers in power conditions of the distribution systems, from the bidding network by using high- behaviour of generators down to precision distribution Phasor residential battery scheduling by Measurement Units (PMUs) small customers through AI/ML/DS techniques. He also applies reinforcement learning to run sample efficient security and stability studies, i.e. to cut down the need to run time-domain simulations.

Page 19 Artificial Intelligence, Machine Learning and Data Science

AI Education & OutreachEnergy

The University of • Corporate advice to support April 2020 by the Queensland interaction between industry Government. Queensland has been partners and UQ researchers, • The AI Hub provides significant successfully operating, aiming to improve their business avenues to raise profile and innovating and engaging • Short courses promote connections between UQ expertise, AI Hub members in the development of • AI focussed undergraduate and postgraduate levels courses and the broader industry, Artificial Intelligence for in computer science, data start-ups and government education and outreach science, information technology, organisations. engineering, mathematics and • Some of the hub activities for many years. Current business analytics. will include advocacy; events outputs include: • The research, research services and forums to increase and education activities around awareness and understanding AI are complemented through of AI applications; and the coordinated community outreach development and delivery of and engagement activities. UQ skills programs, mentoring and Queensland AI Hub training activities. https://www.qldaihub.com is a founding member of the Queensland AI Hub, launched in

Page 20 Artificial Intelligence, Machine Learning and Data Science

Our people

Prof. Amin Abbosh A/Prof. Adrian Athique MR labs. [email protected] [email protected] Prof. Barth’s research interests Prof. Amin Abbosh is the leader of Dr Athique is an expert in cultural include: the Electromagnetic Innovations studies in Asia. • Understanding brain activity team. His main research interests His research interests include digital using functional MRI include: media and society, transnational • Memory consolidation during • Electromagnetic medical imaging media in Asia, technology and sleep and decoding measured systems including hardware economy of media systems, digital functional signals (brain reading) imaging and visual sociology. such as microwave devices and • Ageing and dementia using MR He currently leads a project on digital antennas Neuroimaging to visually detect • Design and analysis of transactions in India. very small venous vessels and microwave devices for wideband Dr Feifei Bai small bleedings in the brain performance, antennas and radio [email protected] • Cardiac MR, at the ultra-high field wave propagation strength of 7 Tesla to examining Dr Bai is an Advance Queensland • Computational electromagnetic, the anatomy and function of the Research Fellow with the UQ School signal processing. human heart. of Information Technology and A/Prof. Udantha Aberyratne Electrical Engineering. Dr Richard Bean [email protected] Her research interests include PV [email protected] integration impacts to power grid, Dr Abeyratne’s research interests Dr Richard Bean is a Research PMU applications in distribution encompass digital signal processing, Fellow at the Centre for Energy Data networks, and fast frequency machine learning, medical Innovation (CEDI). He is an expert in response. instrumentation, medical imaging, the field of data science. electrophysiology, bio-signal analysis Dr Mahsa Baktashmotlagh His research has focussed on areas as and electronics. [email protected] diverse as the combinatory of Latin Dr Abeyrath’s research programs squares and designs, statistics, power Dr Mahsa Baktashmotlagh’s research are characteristic of unorthodox systems analysis (scheduling battery focus is on developing ML and data approaches resulting in pioneering charging using energy forecasting, mining techniques and applying outcomes that produce spin-off and reserve planning in the Australian them to visual data analysis, road companies and patents. NEM), and transport (the effect of traffic networks tools to predict weather and land-use on bike share Dr Mollah Rezaul Alam congestion, prediction of neonatal demand in cities across the world). [email protected] sepsis and hedging foreign exchange trading risks. A/Prof. Pierre Benckendorff Dr Mollah is a Postdoctoral Research [email protected] Fellow. His research expertise Prof. Markus Barth is on characterisation of power [email protected] Dr Pierre Benckendorff has more quality events, machine learning, than 15 years’ experience in education Prof. Markus Barth has an ARC Future , fault detection, and research in the tourism field. Fellowship (2014). He currently classification and analysis of His research interests include visitor leads the ultra-high field human MR distributed energy resources and behaviour, tourism information research program at the Centre for dynamic loads impact. technologies, and tourism education Advanced Imaging. Outcomes from and training. his research on Magnetic Resonance Imaging (MRI) have been used to develop MRI scanner software packages used across the globe in

Page 21 Artificial Intelligence, Machine Learning and Data Science

Our people

Dr Brigid Betz-Stablein • Effectively manage the service ecosystems, and is positioned [email protected] complexity of operations in at the intersection of the Business Dr Brigid Betz-Stablein applies her genomes, and proteomes of Information Systems and Service statistical expertise to support a thousands of dynamically Science disciplines. multi-disciplinary team within the regulated molecules Prof. Robin Burgers-Limerick Centre of Research Excellence for the • Enable the seamless aggregation [email protected] study of naevi, (commonly known as (or integration) of uncertain and Prof. Burgess-Limerick’s research moles), the strongest risk factor of incomplete data, typical of the interests range across the broad melanoma. next wave of biotechnology, scope of human factors and The study uses multiple cameras to across genomics, proteomics, ergonomics from visual perception take simultaneous images of naevi, structural biology, etc. and movement control, through which then are stitched together Dr Steffen Bollman workplace interventions to prevent creating a 3D body avatar. [email protected] injuries due to manual tasks, and It aims to use spatial-temporal Dr Steffen Bollmann’s expertise the design of mining equipment to models to describe the distribution of in magnetic resonances has led reduce injury risks. naevi across the 3D body surface, as to the development of methods well as how naevi change over time. to integrate multiple quantitative Prof. Andrew Burton-Jones Dr Alina Bialkowski imaging modalities to extract hidden [email protected] [email protected] tissue characteristics that could Prof. Burton-Jones’s research focuses Dr Alina Bialkowski is a computer potentially become early biomarkers on how effectively organisations vision and machine-learning for neurodegenerative diseases. use IT (e.g. the effective use of researcher working on applications electronic health records in health in medical imaging and intelligent Dr Edgar Brea authorities); improving methods to transport systems. [email protected] analyse and design IT systems (e.g. Her research interests are in Dr Edgar Brea has significant ways to improve the specification of extracting information from complex experience in the design, user requirements); and improving data and developing interpretable development and implementation of theories and methods for researchers models to solve real-world problems. information technologies, as well as to use in the Information Systems expertise in organisational, sectoral She has developed deep neural discipline. and technology assessments on the networks to understand better impact of innovation at the industry A/Prof. Nicholas Carah human perception and attention in level, and identification of commercial [email protected] driving. and corporate growth opportunities A/Prof. Nicholas Carah is an expert A/Prof. Mikael Boden at the firm level. on digital media and cultures in the [email protected] He is investigating the dynamics School of Communication and Arts. Dr Mikael Boden is a computer of business models innovation and His research focusses on digital scientist with interests in change in technology-intensive and social media platforms, with a computational biology, firms using exploratory data mining particular interest in the interplay bioinformatics, and statistical techniques. between their participatory cultures machine learning. and data-processing power; In bioinformatics, his research aims Dr Christoph Breidbach advertising models; and use of [email protected] to develop and apply methodologies machine vision. to understand and resolve a range Dr Christoph Breidbach’s empirical of open problems in genomics, and conceptual research addresses molecular and systems biology, the fundamental question of how including how to: information technology transforms

Page 22 Artificial Intelligence, Machine Learning and Data Science

Our people

Dr Shakes Chandra methods) and phenotyping (aerial The commercial and academic [email protected] imaging, canopy monitoring, remote impact of his work in Magnetic Dr Chandra’s expertise includes sensing) to understand better the Resonance Imaging has been image processing, machine learning, interactions between the biophysical significant, with about two-thirds discrete tomography and medical environment and genetics, growth of all high-end, clinical MRI systems image analysis. He developed and development on crop yield. installed worldwide after 1997 deformable model and machine- In recent years, Prof Chapman’s containing patented technology co- learning algorithms for knee research has linked with IT, invented and fully developed by him. deployment on the Siemens Syngo engineering and maths experts at Dr Yi Cui platform. His research interests are: UQ to utilise digital agriculture to assist plant breeding programs. [email protected] • Machine/Deep learning He uses ML, deep learning and Dr Cui is a Research Fellow in the UQ • Making MRI faster and more imaging processing tools to improve School of Information Technology affordable through better image plant phenotyping and understand and Electrical Engineering. His reconstruction, processing and its impact on productivity. research interests include wide- analysis Some of his current projects include: area monitoring and control, data • Applying fractals, chaos, • INVITA, analytics and machine learning of and number theory to image distribution networks, condition • An international collaboration to processing and computer assessment and fault diagnosis of develop improved pipelines for science. power transformers. deep learning analysis of farm Dr Archie Chapman crop images. Dr Alan Davidson [email protected] • Use of climate forecasting and [email protected] Dr Chapman develops and applies satellite remote sensing to Dr Alan Davidson is a lawyer and a principled AI, game theory, analyse crop types and forecast senior lecturer in the UQ School of optimisation and ML methods to national wheat yields Business. His expertise is on the legal solve large-scale and dynamic aspects of: allocation, scheduling and queuing Dr Cedric Courtois • International trade finance problems. [email protected] • Electronic commerce His recent research has focused on Cedric Courtois studies structural applications of these techniques to and agentic properties of (social) • Technology problems in future power systems, media-related audience practices • Privacy such as using batteries and flexible and their consequences in a variety • International banking loads to provide power network of contexts (e.g. the workplace, • Artificial Intelligence and system services, while making intimate relations, and civic the best use of legacy network and engagement/and or participation). A/Prof. Gianluca Demartini generation infrastructure. His research interests include [email protected] algorithms, online platforms and Dr Demartini’s research interests Prof Scott Chapman audience research. [email protected] include information retrieval, Scott Chapman is Professor in Crop Prof. Stuart Crozier , and human Physiology. He studies genetic [email protected] computation. and environmental effects on the Prof. Stuart Crozier’s research His work focuses on improving physiology of field crops, particularly expertise lies in imaging technology the efficiency and effectiveness where drought dominates. He and applications, instrumentation of human-in-the-loop artificial applies quantitative approaches for physiological measurement and intelligence systems. (crop simulation and statistical semi-automated diagnostics. The application domains of his

Page 23 Artificial Intelligence, Machine Learning and Data Science

Our people

research include text analytics and Dr Chandima Ekanayake Prof. Nicole Gillespie the intersection between structured [email protected] [email protected] and unstructured digital content. Dr Ekanayake’s research experience Prof. Nicole Gillespie’s research He has collaborated with several and interests include condition focuses on trust development and industry and governmental monitoring of power apparatus, repair in contexts where trust is organisations including, Facebook, alternatives for insulating oil, challenged (e.g. after a trust failure, in Google, Microsoft, Yahoo, IBM, SAP, performance studies of High Voltage complex stakeholder environments, and The National Archives in the UK. insulators and energy related themes. during organisational transformation and digital disruption, in virtual Dr Arindam Dey A/Prof. Anders Eriksson healthcare and Artificial Intelligence, [email protected] [email protected] in cross-cultural relations). Dr Dey’s research interests are in A/Prof. Eriksson is an Australian Current research projects focus on designing novel interfaces using Research Council Future Fellow. His understanding stakeholder trust both Artificial Reality, Virtual Reality, research areas include optimisation in organisations and industries, VR, and a combination of both (Mix theory and numerical methods organisational trust repair, designing Reality) and their application in applied to the fields of computer trustworthy organisations, trust in multiple domains, including affective vision and machine learning. Artificial Intelligence, and stakeholder computing, gaming, education, trust and uptake of telemedicine. health, and sustainability. Prof Michael Forbes Michael Forbes is a Senior Lecturer He is also interested in Empathic Dr Mashhuda Glencross at the School of Mathematics and Computing, a research field that [email protected] Physics. Michael does research in develops computer systems that Dr Mashhuda Glencross research Operations Research and applied recognise and share emotions and interests are in computer graphics mathematics. His expertise covers help people to better understand one and human-computer interaction and algorithms, machine learning, linear another. include virtual/augmented reality, programming, optimisation, routing IoT, modelling the reflectance and and data mining, to name but a few. Dr Chelsea Dobbins structure of materials for high-quality [email protected] graphics rendering and 3D scene A/Prof. Marcus Gallagher Dr Chelsea Dobbins is an expert on [email protected] reconstruction. digital health and human-computer interaction. Dr Marcus Gallagher’s main Prof. Geoffrey Goodhill [email protected] Her research focuses on the research interests are metaheuristic detection of emotions using optimisation and machine learning Prof. Goodhill leads a multidisciplinary smartphones/wearable sensors and algorithms, in particular techniques team working on how brains process personal informatics. based on statistical modelling. information, particularly during development. This includes how This includes areas such as He is also interested in biologically growing nerve fibres use molecular lifelogging, pervasive computing, inspired algorithms, methodologies cues to make guidance decisions, digital health, human-computer for empirical evaluation of algorithms how map-like representations of interaction, ML, mobile computing, and the visualisation of high- visual inputs form in the optic tectum mobile/wearable sensors, human dimensional data. and visual cortex, and how these digital memories, signal processing, maps code sensory information. and physiological computing. He addresses these questions using a combination of mathematical, experimental and computational techniques.

Page 24 Artificial Intelligence, Machine Learning and Data Science

Our people

Prof. Graeme Hammer projects which have successfully analysis (event detection and [email protected] implemented genomic technologies monitoring; user behaviour modelling Prof. Graeme Hammer researches the in livestock and cropping industries. and prediction; recommender physiology and genetics of complex systems) and big database Prof. Paul Henman adaptive traits in field crops with management. [email protected] a focus on water productivity in cereals. Prof. Paul Henman’s research focuses Dr Stan Karanasios His research underpins the on the relationship between social [email protected] development of mathematical policy, administration and digital Dr Stan Karanasios’s research focus models of crop growth, development information technologies. on how organisations use, adapt, and and yield that enable simulation This space includes welfare state navigate new digital technologies. of consequences of genetic and and reform; e-government; and the He is currently working with a management manipulation of crops administration of policies, among leading European car manufacturer in specific target environments. other. to examine ways of re-imagining A/Prof. Jim Hanan Increasingly algorithmic government work process to accommodate an [email protected] and AI is heightening concerns with increasingly digital product. government accountability and Dr Jim Hanan has developed transparency, and reproduction of mathematical, computational bias, discrimination and inequalities. and visualisation approaches and techniques to facilitate the study of Prof. Paul W Hodges genetics, physiology, morphogenesis [email protected] and ecology at the scale of cells and individual plants and insects. Paul Hodges is recognised as a Prof. Michael Haugh world leader in movement control, [email protected] pain and rehabilitation. His unique comprehensive research approach Michael Haugh is Professor of from molecular biology to brain Linguistics. His research is primarily physiology and human function in the field of pragmatics, with has led to discoveries that have a focus to date on analysing transformed understanding of face, (im) politeness, teasing and why people move differently in humour, indirectness, and intention. pain. His innovative research has He works with recordings and also led to discoveries of changes transcriptions of naturally occurring in neuromuscular function across spoken interactions, as well as data a diverse range of conditions from from digitally mediated forms of incontinence to breathing disorders. communications, across several languages. A/Prof. Helen Huang Prof. Ben Hayes [email protected] [email protected] Dr Huang’s research interests mainly Prof. Ben Hayes’ researh is in genetic include multimedia search (content improvement of livestock, crop, understanding and analysis; video pasture and aquaculture species, with and image captioning; multimedia a focus on the integration of genomic indesing and retrieval; near- information into breeding programs, duplicate detection; ranking and including leading many large-scale recommendation); social media

Page 25 Artificial Intelligence, Machine Learning and Data Science

Our people

Dr Erkan Kayacan Dr Jiwon Kim Prof. Dirk Kroese [email protected] [email protected] [email protected] Dr Erkan Kayacan is a lecturer in the Dr Kim’s research is broadly in the Prof. Dirk Kroese’s AI research UQ School of Mechanical Engineering. area of modelling and analysis involves the understanding and He has been involved in the of intelligent transport systems use of randomised (Monte Carlo) development and commercialisation (ITS) and human mobility, with an methods to help solve difficult of several robotic systems and was emphasis on vehicle trajectories estimation and optimisation recipient of the Best Systems Paper in large-scale traffic networks and problems in Data Science and ML. Award at Robotics: Science and predictive analytics for real-time He is a pioneer of the well-known Systems (RSS) in 2018. traffic management and operations. Cross-Entropy method. His research focuses on discovering Dr Hassan Khosravi patterns and insights from massive Prof Paul Lever [email protected] collections of urban movement data; [email protected] Dr Hassan Khosravi’s research draws using data mining and ML algorithms Prof. Lever’s main research has on theoretical insights from learning to develop solutions to transport- been the theory and application of science and exemplary techniques related challenges in our cities. intelligent systems to earth science from human-computer interaction, applications, with a particular learning analytics and explainable AI, Prof. Peter Knights emphasis on intelligent robotic and to design, implement, validate and [email protected] automated systems for dynamic deliver technological solutions to Prof. Peter Knights’ research environments. Other research learner-centred, data-driven learning interests are in mine maintenance interests include mine automation at scale delivery. management, mine operations and and robotics, development and process control, mining simulation, implementation of software based A/Prof. Dan (Dongseong) Kim mining equipment automation. on AI, and Computer Based [email protected] His recent work has focussed on Management Information and Dr Dan Dongseong’s research mine data analytics and the reliability Intelligent Decision Support Systems interests are in cybersecurity and and planning issues associated (among others). dependability for various systems with novel mining systems. and networks. Prof. Xue Li [email protected] Dr Ryan Ko [email protected] Dr Xu is an expert in data. His research interests are in data mining, Prof. Ryan Ko is Chair and Director of intelligent information systems, and UQ Cyber Security at the University social computing. The Australian of Queensland, Australia. Financial Review recognised him as His research reduces users’ reliance one of the top 50 most powerful on trusted third parties and people in Australia in 2015. focusses on provenance logging and reconstruction, traceability and privacy-preserving data processing (homomorphic encryption)

Page 26 Artificial Intelligence, Machine Learning and Data Science

Our people

Prof Feng Liu Dr Ben Matthews paradox that such classifiers can [email protected] [email protected] actually have lower error rates than those formed from a completely Prof. Feng Liu’s research lies in Dr Matthews’s research focuses on classified sample in certain cases. medical imaging, with the focus the human aspects in the design on Magnetic Resonance Imaging and use of technology; the methods Dr Alejandro Melendez-Calderon hardware design and electromagnetic for studying it; and the involvement [email protected] analysis, and cardiac electrical of people in the design processes function imaging. (human-computer interaction). He Dr Alejandro Melendez-Calderon has an interdisciplinary background in His current research program has worked in a range of design robotics and biomedical engineering, includes MRI reconstruction (parallel domains with various industry with extensive experience in human imaging, compressed sensing, etc.), partners imcluding: audiology augmentation technologies used in Bioelectromagnetism, ML and Deep (Oticon), diabetes care (Novo medicine (robotics, wearable devices) Learning, among others. Nordisk), domestic devices, industrial components and computational approaches to Prof. Brian Lovell (Danfoss). understand human neuromuscular [email protected] control (unimpaired, stroke and SCI Prof. Ross McAree population). Prof. Lovell is an expert in computer [email protected] Dr Melendez-Calderon’s research vision, machine-learning and pattern aims to understand the principled recognition methods. His research Prof. McAree’s research focuses on mechanisms of human behavior, interests are in: machinery dynamics and control with a current emphasis on mining in particular related to movement • Hidden Markov Model Learning equipment automation. Prof. control/learning and physical Theory McAree led the development and interaction; his technical interests • Technology demonstration of the world’s first are in robotics and computational • Face Recognition fully autonomous mining excavator modeling for medical diagnostics, assistive applications and biomedical • Image Segmentation with as the outcome of a 10-year education. Geodesic Level Sets collaboration with Joy Global Surface Mining (now part of Komatsu). He Dr Lakshitha Naranpanawe Dr Hui Ma has also collaborated with Caterpillar [email protected] [email protected] Inc. to develop and trial the world’s Dr Ma’s research interests include first autonomous bulldozer capable Dr Lakshitha is a Postdoctoral industrial informatics, power systems, of production dozing using the ‘pivot Research Fellow at the UQ School high voltage engineering, wireless push’ method. of Information Technology and sensor networks, and sensor signal Prof. Geoffrey McLachlan Electrical Engineering. His research processing. [email protected] focuses on power transformers, modelling, condition monitoring. Prof. Greg Marston Prof. Geoffrey McLachlan’s research [email protected] interests are in areas associated A/Prof. Yoni Nazarathy [email protected] Prof. Greg Marston’s research with the statistical machine learning interests are on the social impact of aspects of artificial intelligence. Dr Nazarathy specialises in stochastic AI and automation in the future of The most recent activities of his operations research, scheduling, work and public policy approaches research team have centred on the optimisation, control and statistics. to regulating AI to reduce bias and performance of classifiers formed His expertise can be applied to increase transparency in algorithmic from partially classified data. agriculture, power engineering and decision-making. They have produced a number of health context. seminal results including the apparent

Page 27 Artificial Intelligence, Machine Learning and Data Science

Our people

Dr Quan Nguyen computational engineering and unstructured data to help machines [email protected] mechanics. understand the conversational language in emergency situations Dr Quan Nguyen leads the Genomics A/Prof. Marius Portmann and health informatics. and Machine Learning (GML) [email protected] unit at the Institute for Molecular Dr Portmann’s research interests are Prof. Aleksandar Rakić Bioscience (IMB). The GML lab has [email protected] been a pioneer in developing neural in the area of Computer Networks network models for connecting tissue (Wireless Mesh Networks, Network Aleksandar D. Rakić leads morphology with molecular gene protocols, Software Defined UQ’s Photonics and Microwave expression data for discovering in Networking, P2P Computing), Cyber Engineering group, focusing on situ biological processes underlying Security, and blockchain technology. the development of technologies tissue heterogeneity, focusing on for sensing and imaging across the cancer and genetic disease. His Dr Andries Potgieter electromagnetic spectrum including [email protected] research uses multidisciplinary microwave, terahertz wave and expertise in bioinformatics, genomics, Dr Potgieter leads a team of optical systems. systems biology, biostatistics,and researchers in the areas of seasonal The Rakić group has pioneered machine learning to explore cutting- climate forecasting at a regional scale, development of several world edge topics in genomics and remote and proximal sensing with firsts, including laser-feedback transcriptomics, including analysis of applications in the development of interferometric sensors, methods some of the world’s largest genomics crop production outlooks and less based on monolithic Vertical-Cavity data sets. risk-prone cropping systems across Surface-Emitting Laser arrays Australia. (VCSELs), blue-green lasers, terahertz Dr Antonio Padilha quantum cascade lasers and mid- [email protected] Dr Amelia Radke infrared interband cascade lasers. [email protected] During the past decade, Dr Padilha has been engaged in developing Dr Amelia Radke’s research lies at the technology for the healthcare sector, juncture of science and technology, Dr Tapani Rinta-Kahila particularly the design of systems for anthropology, and socio-legal [email protected] patients in rehabilitation or assistive studies. Dr Rinta-Kahila is a Postdoctoral settings. Dr Radke has developed policy Research Fellow working in The process involved an integration of submissions on evaluation strategies the Digital Enterprise research neuro engineering techniques, robotics, for programs affecting Aboriginal program at the Australian Institute control, and instrumentation to create and Torres Strait Islander peoples, for Business and Economics. devices and algorithms to sense and along with the application of a His research addresses the control human motion in real-time. human rights approach to emerging organisational implementation of He aims to develop and evaluate technologies (AI-driven) in Australia. artificial intelligence and automation, non-invasive technology to improve the identification and prevention of Dr Ash Rahimi rehabilitation for people with motor technologies’ negative effects on [email protected] disabilities. individuals and organisations, as well Dr Rahimi’s research interests fall as the discontinuance of incumbent Dr Dorival Pedroso within the fields of Natural Language information technologies. [email protected] Processing, Social Network Analysis Dr Pedroso is an expert in and Machine Learning. numerical and computer methods She is specifically interested in for solid mechanics and materials exploiting both structured and modelling. His research focuses on

Page 28 Artificial Intelligence, Machine Learning and Data Science

Our people

Prof. Daniel Rodriguez Dr Peter Scarth Dr Carl Smith [email protected] [email protected] [email protected]

Prof. Daniel Rodrigues is a leader Dr Scarth has broad remote sensing Dr Smith has expertise in in agricultural systems modelling, skills across terrestrial and aquatic Bayesian Networks, land resource whole-farm resource allocation. environments. His works with TERN/ management, geographic Prof. Rodriguez’s team applies Auscover aims to democratise spatial information systems, remote systems thinking and analysis tools to data access and use, and delivering sensing, systems modelling crop systems. validated national and global and decision support systems. scale products for friendly use by Dr Fred Roosta-Khorasani scientists, policy and the public. Prof Simon Smith [email protected] [email protected] Prof. Graeme Shanks Dr Fred Roosta-Khorasani is an ARC Wrist-worn devices (Garmin, Fitbit, [email protected] Fellow at the School of Mathematics apple watch, etc.) have become and Physics. His research focuses Prof. Graeme Shanks works in the popular and can provide a high on the design, analysis, and UQ School of Business. His research volume of meaningful data about implementation of novel optimisation interests include ethical aspects of everyday routines. algorithms for training modern ML big data analytics, understanding Prof. Smith’s research team uses AI/ problems. how organisations benefit from AI, ML approaches to classify high- business analytics and enterprise frequency activities (e.g. on-road Prof. Shazia Sadiq architecture. driving) from wrist-worn (actigraphy) [email protected] devices aiming to bridge the Dr Rahul Sharma Shazia Sadiq is part of the Data biological and social sciences for [email protected] Science research group, Her research community benefit. interests are innovative solutions for Dr Sharma research focuses on

Business Information Systems that control systems applications. This Dr Ida Someh [email protected] span several areas including business encompasses system modelling, process management, governance, control development and model- Dr Ida Asadi Someh is a research risk and compliance, and information based fault diagnosis. affiliate at the Centre for Information quality and use. He applies his research to solar farm Systems Research (CISR). Her research focuses on organisational Prof. Tapan Saha fault detection and diagnosis, control and societal impact of data, analytics [email protected] of grid-connected inverters, and control algorithms for demand-side and artificial intelligence. Prof. Saha is an expert in renewable management. energy. Some of his current research A/Prof. Clair Sullivan projects include: Dr Sally Shrapnel [email protected] [email protected] • Solar/wind/geothermal energy Dr Sullivan leads the Queensland integration to electricity grid Dr Shrapnel works at the interface of Digital Health Research Group. She • Power systems analysis- causality and ML. Her goal is to infer is a leading Australian researcher renewable energy integration causal relationships from different on electronic medical records and (Solar PV and wind) types of data, and build models the digital transformation of health that are robust with respect to • Intelligent Diagnostics of Aged systems. interventional distributional shifts. Power Transformers and other

ageing assests Her research combines theory, methodology, and applications across • Electricity market analysis. a broad range of disciplines.

Page 29 Artificial Intelligence, Machine Learning and Data Science

Our people

Dr Hongfu Sun modelling and signal’s processing to designing solutions tailor to [email protected] paediatric and adult respiratory and the needs of individuals who sleep medicine. undergoing a rehabilitation Dr Sun’s research interests include journey developing novel MRI contrast Prof. Juha Toyras • Design, development and mechanisms, e.g. Quantitative [email protected] evaluation of technological Susceptibility Mapping (QSM), fast Juha Toyras is a Professor in interventions to reduce fuel and multi-parametric MR acquisitions, Biomedical Engineering. His research consumption and improve safe as well as advanced image analysis is focused on medical devices and driving associated with pollution techniques, e.g. machine/deep sensors, medical imagining, medical and road trauma impacts on learning, to study neuroscience and signal processing and connective human health. neurological diseases. tissue biomechanics. His work has led to patents and commercialization Dr Slava Vaisman Dr Thomas Taimre of scientific innovations including [email protected] [email protected] a forehead EEG-electrode Dr Radislav Vaisman’s research Dr Thomas Taimre current research set for emergency medicine. interests lie at the intersection of expertise lies in probability applied probability, statistics, and theory, computer simulation, and A/Prof. Mark Utting computer science and its application mathematical optimisation with [email protected] to theoretical and real-life problems, biological and other scientific, Dr Mark Utting’s research interests in the fields of machine learning, engineering, and finance disciplines. include software verification and AI/ optimisation, safety, and system model-based testing, and ML. Dr Gayani Tennakoon reliability research, among others. Mudiyanselage He is passionate about designing Dr Vaisman has made key [email protected] and engineering good software that contributions to the theory and solves real-world problems, and has the practical usage of Sequential Dr Tennakoon Mudiyanselage’s extensive experience developing Monte Carlo methods specifically to research involves using health and next-generation genomics software the development of the stochastic administrative data to developing and manufacturing software. enumeration method for estimating explainable ML learning models to the size of backtrack trees. support decision-making. Dr Atiyeh Vaezipour She has applied unsupervised [email protected] Prof. Rhema Vaithianathan ML methods to develop effective Dr Atiyeh Vaezipour’s background [email protected] approaches to discover knowledge is in Computer Interaction (HCI)/ Prof. Rhema Vaithianathan is an from social media networks based on User Experience (UX) in health and expert in data analytics for social user interaction patterns. road safety research settings. Dr good. She has a strong interest in the Vaezipour’s is part of a research team Dr Philip Terrill use of predictive analytics to support aiming to: [email protected] decision making in health and human • Develop and evaluate innovative services. Dr Terrill’s research interest is in rehabilitation technology Prof Vaithianathan has collaborated the development of novel medical solutions to improve the delivery with hospitals and health policy diagnostic tools and therapies to and availability of evidence- agencies for more than 25 years. improve the health outcomes of based health services people in Australia and globally. • Enhance user experience and His current research focusses improve the interaction between on the application of electronic end-users and technology by instrumentation, mathematical

Page 30 Artificial Intelligence, Machine Learning and Data Science

Our people

A/Prof. Eric J. Vanman Prof. Janet Wiles Queensland and Noja Power, Powerlink, [email protected] [email protected] and Redback (among others). Dr Vanman’s research interests Prof. Wiles is an expert in language. Dr Nan Ye include the social neuroscience Her current research projects [email protected] aspect of emotion and intergroup focus on human-robot interactions, Dr Nan Ye is broadly interested in prejudice. He applies several language technologies, bio-inspired algorithms, theory, and applications kinds of psychophysiological and computation, visualisation and that involves learning from data for neuroimaging methods in his studies. artificial intelligence. automated decision-making. Dr Vanman’s work on Previous projects have been in His research has been applied to unconscious bias displayed via complex systems modelling in autonomous driving, automated error psychophysiological measures was biology and neuroscience, human detection and correction of survey among few early studies that laid the memory, language and cognition. data, species distribution modelling, groundwork for current research on Dr Ian Wood and information extraction. implicit measures. [email protected] He has expertise on: Recently, he has used a social • Probabilistic graphical models neuroscience approach to study the Dr Ian Wood’s research interests mechanisms of empathy, including are in classification, bioinformatics, • Planning under uncertainty factors that might lead to a failure of stochastic optimisation, machine (focusing on partially observable empathy for others who are different learning and mixture models. Markov decision processes) to us. His new line of research • Active learning A/Prof. Mohsen Yahyaei examines how we really feel about [email protected] • Inference and learning with robots. Can we have empathy for complex losses robots? Why do we fear them? Is it a Dr Mohsen Yahyaei is leading the good idea to design robots that look Advanced Process Prediction and Dr Hongzhi Yin [email protected] like humans? Control (PPCo) program aiming to transform unit process modelling and Dr Yin’s main research interests Dr Sen Wang simulation, moving from the steady- include the recommender system, [email protected] state models. social media analytics and mining, Dr Sen Wang’s research Iinterests The program will develop and network embedding and mining, time include medical data analysis, signal apply techniques to make greater series data and sequence data mining and image processing, pattern use of data generated on-site and and learning, chatbots, federated recognition and machine learning sensor technologies in combination learning, topic models, deep learning algorithms. with advanced process control, and smart transportation. computational analytics and He is the current director of RSBDI modelling techniques. (Responsible and Sustainable Big Dr Richard Yan Data Intelligence) Lab. [email protected] RSBDI Lab aims and strives to develop energy-efficient, privacy- Dr Yan is an ARC DECRA Fellow. preserving, robust, explainable His expertise includes distributed and fair data mining and machine renewable energy integration to learning techniques with theoretical power grids, big data mining in power backbones to discover better systems, and network strength and actionable patterns and intelligence He has collaborations with several from large-scale, heterogeneous, industry partners including, Energy networked, dynamic and sparse data.

Page 31 Artificial Intelligence, Machine Learning and Data Science

Our people

Dr Dongming Xu Prof. Xiaofang Zhou [email protected] [email protected] Dr Xu’s research focuses on the Prof. Xiaofang Zhou’s research understanding of how information interests are in big data systems are used and how analytics, spatial and multimedia information systems influence databases, information system society. integration, data quality She usually combines theoretical management, and high- model building, laboratory and field performance query processing. experiments to develop prototype His research focusses on finding systems. She works on the use of effective and efficient solutions social media in business and society. to managing, integrating and She is also interested in the analysing a large amount of theoretical foundations and complex data for business and applications of AI to decision-making scientific applications. and business intelligence. In recent Prof. Guido Zuccon years, she has focused on the hi-tech [email protected] start-ups development. Prof. Zuccon leads the Information Engineering Lab Dr Miao Xu (ielab), a research team working [email protected] in Information Retrieval and Dr Miao Xu’s research focuses on Health Data Science. machine learning and its real-world Guido’s main research interests applications. are formal models of Information She has applied the proposed Retrieval, Health Search, Formal, methods to bioinformatics and Models of Search, and Health computer vision problems. Data Science. Prof. Zuccon Her long-term goal is to develop has studied AI/ML methods to intelligent systems, which can learn estimate the understandability from a massive volume of complex of health web pages and use (e.g. weakly supervised, incomplete, these to improve the retrieval of noisy) data (e.g. single-/multi- information for people seeking labelled, ranking). health advice on the web.

Page 32 Artificial Intelligence, Machine Learning and Data Science

If you would like to know more about UQ capability in these areas or how to engage with the appropriate reseacher/s please contact:

Professor Shazia Sadiq Deputy Head School of Information Technology and Electrical Engineering [email protected]

Professor Mohan Krishnamoorthy Pro-Vice-Chancellor (Research Partnerships) [email protected]

Page 33 Research Partnerships Office of Deputy Vice-Chancellor (Research and Innovation) The University of Queensland Australia T +61 7 3365 3559 [email protected]

March 2021