- understanding and enabling a smart data-driven world and
- creating future digital technologies.
D2I brings together advanced data and computer scientists, mathematicians and engineers. Our research is multi-disciplinary and solves both the industrial and the theoretical research problems. Our aim is to build advanced and automated data- and process-driven models and intelligent algorithms to solve real-world problems. As all datasets and processes are different, we design tailor-made solutions for individual problems, taking into account new hardware and the underpinning computational platforms. We model complex systems and processes and develop multi-level and multi-modal algorithms.
We integrate machine intelligence with mathematical and physical modelling, and optimal and multi-criteria decision making to find knowledge from data. We use advanced data science technologies, theoretical approaches, advanced computational methods, optimisation and mathematical modelling for automatic data analysis, machine learning, planning and decision making to solve problems in the domains of health, food, agriculture, defence, energy, environment, sport, travel, finance and education.
We have strong and unique expertise that combines all aspects of data, computational and mathematical predictive modelling, powerful intelligent algorithms for knowledge discovery and optimal multi-criteria decision making. The centre is one of the largest in the School of IT. Our expertise is diverse, extensive and complementary with an excellent track record in multi-disciplinary projects. We have strong publications in journals and top conferences. Our research is funded externally with several ARC-funded projects, as well as defence and industrial funding.
The following are our expertise areas:
Mathematically approximate unknown functions to minimise the margin of error, with applications in real life systems, such as nuclear physics and computing
Advanced multicriteria decision making methods based on cooperative games, nonlinear integrals and fuzzy systems
Scheduling and Planning
Providing solutions to large-scale complex scheduling problems, such as timetabling, resource allocation, scheduling in the manufacturing industry, facility location planning, and supply chain and logistics
Developing advanced mathematical and computational models based on dynamical systems and differential equations with applications to physiology, health, medicine and defence.
Using advanced methods for nonlinear and dynamical signal processing and time series analysis for data driven modelling of human behaviour, processes of sleep, cardio-respiration, diabetes, mental health, age-related conditions.
Mining data, recognising patterns, big data dimensionality reduction and visualisation, clustering and classifying data, modelling interactions and causality to better understand systems and behaviours
Developing intelligent solutions for data compression, data partitioning, data farming, multi-source data fusion for making smart the way we live and the way we keep healthy through smart homes and eHealth.
Using and developing complex models and algorithms for supervised and unsupervised machine learning, and machine learning driven knowledge system development and management to make accurate predictions and obtain meaningful insights from data
Anomaly Detection, Prediction and Management
Detecting meaningful anomalies from streamed data, including social media data, data collected from IoT devices and video data.
Providing efficient algorithms for detecting and recognising humans, objects, devices, scenes, handwriting, with advanced biometrics, provenance attribution, scene analysis, video surveillance techniques in security and intelligent processes and for decision making.
Aiming to leverage information about the context of acquisition and the image formation process to develop methods, algorithms and techniques for processing visual and multimodal data requiring the least amount of user intervention possible.
Human-aligned Machine Learning
Investigating and developing approaches to ensure AI systems behaviour remains beneficial to humanity concentrating on three foci: ethics, safety, and interactivity (explainability, instructability).
Natural Language Processing (NLP) for detecting disinformation and finding patterns in data, text retrieval, text classification, textual entailment analysis, natural language question answering, and conversational question generation using sequence-to-sequence and attention neural models.
Developing index techniques and algorithms to manage traffic network data and supporting advanced city planning and scheduling in the complex scenarios;
Developing the novel social computing models and algorithms for Innovating the online advertisement strategies and tracking the evolving change of social media trends;
Designing the AI-powered modern database system to build the bio-information knowledge graph and innovate the lifecycle of health ecosystem.
Rigorous privacy models to guarantee the bounds of unintended disclosure individual’s data, and designing privacy preserving techniques for various data science tasks and AI methods.