Models and algorithms

Program vision

Program Leader: Professor Chris Drovandi

Program Co-Leaders: Professor Timothy Moroney and Professor Matthew Simpson

The underlying laws and relationships driving complex real-world systems can be unknown, but manifest as structure and patterns within observed data. Whilst more data is being collected and stored than ever before, there is no value in data until it is converted to information and decisions. Data focused models and algorithms bridge the gaps between data, information and decisions, providing a conduit to extract value from data.

The Centre will consolidate know-how within the university, both in fundamental methods and domain focused applications in data focused modelling. This collective along with our industry, university and international partners will provide direction, capability and capacity within the program to address fundamental challenges in data focused modelling. This research program will develop new approaches to discovering models which describe hidden structure in data, and explore new ways to exploit these models.

Often complex models are required to capture the intricacies of real processes. This motivates the need for computational algorithms to calibrate them.  Efficient and accurate computational algorithms are central to many data science tasks such as model selection, experimental design, optimisation, clustering and prediction.  This program will consolidate and create new and accessible computational algorithms that are faster and more accurate to enable new discoveries across science and technology, addressing modern obstacles such as complex mathematical models and challenging data. This will reduce the burden on HPC and cloud computing resources. We will achieve this by advancing and melding statistical, machine learning and computer science algorithms that are capable of exploiting modern parallel computing architectures.

Through open dissemination of this work, including open source software and tutorials, data science practitioners and researchers internationally will benefit from our discoveries. This program also has obvious links to all programs in the Centre for Data Science.  It will support innovation in data acquisition, process data-focused models, complement new data science methods and facilitate decision making under uncertainty in a variety of disciplines.