In partnership with a world-leading robotics company, the UBTECH Sydney Artificial Intelligence Centre brings together a multidisciplinary team of dedicated researchers to explore new horizons in artificial intelligence (AI).
The UBTECH Sydney Artificial Intelligence Centre is committed to advancing AI to endow machines with the capabilities of perceiving, learning, reasoning and behaviour. Our researchers design shallow or deep models to extract, represent and understand information encoded in data and build algorithms and theories.
We aim to establish, analyse and evaluate models that can: learn and make predictions on data; create prototypes or applications to investigate autonomous agent actions; and identify patterns and apply logic. Ultimately our vision is to lead AI research in Australia and become one of the most prestigious AI research hubs in the world.
AI is a transformative technology that promises tremendous societal and economic benefit. It has the potential to revolutionise how we live, work, learn, discover and communicate. Research into AI can advance Australia's national priorities, including greater economic prosperity, improved educational opportunities and quality of life, and enhanced national and homeland security.
Australian Research Council Laureate Fellowship Project (2018–2022)
The project aims to develop a suite of original models and algorithms for processing and understanding videos captured by moving cameras, and to establish the mathematical foundations for deep learning-based computer vision to provide theoretical underpinnings. We expect to generate new knowledge that will transform moving-camera computer vision with step changes in visual quality enhancement, compression and acceleration technologies, and solutions for fundamental computer vision tasks. A new concept of feature complexity for measuring the discriminant and learnable abilities of features from deep models will also be defined. The outcomes of the project will be critical for enabling autonomous machines to perceive and interact with the environment.
Australian Research Council Discovery Project (2014–2017)
Our partners: Dr Jun Li (UTS); Professor John Shawe-Taylor (University College London)
This project will develop nonlinear transfer distance metric learning algorithms for training and test samples that are not independent and identically distributed, or from different instance spaces. New theoretical foundations for crowdsourcing will lead to innovative intelligent systems for such purposes as the NBN, social services and security services, and keep pace with developments in hardware technology. Outcomes include applications in social networks, the internet and climate change, as well as video surveillance to help combat crime and terrorism. The innovative research will benefit Australia's economy, environment and society substantially, and will maintain Australia's world-leading role in machine learning and computer vision.
Australian Research Council Future Fellowships (2013–2017)
Data analytics in video surveillance and social computing is a problem because data is represented by multiple heterogeneous features. This project will develop a Multiview Complete Space Learning framework to exploit heterogeneous properties to represent images obtained from sparse camera networks. It will integrate multiple features to identify people and understand behaviour, to build a database of activities occurring in a wide area of surveillance. It will expand frontier technologies and safeguard Australia by providing warnings for hazardous situations (for example, overcrowding and trespassing), as well as criminal and terrorist activity. Results will be applicable internationally and enhance Australia's role in machine learning and computer vision communities.
Australian Research Council Linkage Infrastructure, Equipment and Facilities Projects (2014)
Our partners: Professor David Abramson (UQ), Professor Xiaofang Zhou (UQ), Professor Debra J Bernhardt (UQ), Professor Chengqi Zhang (UTS), Professor Xingquan Zhu (Florida Atlantic University USA)
The 21st century has been described as the century of data. Experts predict an exponential growth in the amount of data that will be captured, generated and archived. Australia has made significant progress towards addressing some of the opportunities and infrastructure challenges posed by such rapid increase in data volumes. However, these investments do not address the growing need to process data. Conventional supercomputers are unable to meet the challenges of the data explosion. The large gap in latency and bandwidth between the processor, memory and disk subsystems means the processor is often idle waiting to fetch data. This LIEF proposal will build a platform focused on data-intensive science.
Australian Research Council Discovery Projects (2012–2014)
Smart information use is essential for effective video surveillance to guard against accidents, fight crime and combat terrorism. In this project, advanced probabilistic methods will be applied to visual surveillance information to warn of impending accidents and to track criminals and terrorists and predict their behaviours.
Australian Research Council Linkage Projects (2015-2018)
Our partners: Associate Professor Ivor Tsang (UTS), Associate Professor Xue Li (UQ), Dr Guodong Long (UTS)
This project plans to build an interactive mining system to detect cyberbullying on social networks that have a large number of participants and a variety of inputs, including conversation texts, time-variant changes and user profiles. The project is designed to change the existing cyberbullying prevention services from reactive keyword filtering to proactive social interaction pattern mining. The intended outcome will enable the early detection and warning of cyberbullying and approach a new way to discover interaction patterns with a large number of participants over evolving and complex social networks.
Multi-object tracking is challenging because objects may cross and occlude each other while moving. When an object is occluded, it is difficult to track its position and, as it reappears, the challenge is to find the correspondence between the visible object and the previously tracked object. We have devised a semi-online model to describe the relationships between the trackers and the bounding boxes in a sequence of frames. Temporal constraints are enforced to ensure the temporal continuity, while spatial constraints are enforced to guarantee the separability of different persons. Here is one tracking example. The proposed tracker obtained exceptional performance on the 2016 multi-object tracking challenge with the public detector.
We propose a coarse-fine network composed of several coarse-detector branches – each of which is built on top of a feature layer in a convolutional neural network – and a fine-detector branch built on top of multiple feature layers. We supervise each branch by a specified label map to explicate a certain supervision strictness level. All the branches are unified principally to produce the final accurate key point locations. We demonstrate the efficacy, efficiency and generality of our method on several benchmarks for multiple tasks including bird part localisation and human body pose estimation. Here is an example.
Our centre offers researchers a world-class education, a great opportunity to work on cutting-edge projects, and a catapult for your career. You will work on projects to solve real-world problems that will help improve people’s lives while being mentored through the PhD program that involves both academic and industry training as well as collaboration with other researchers working on similar problems.
If you are interested in the fields of computer vision, deep learning, data mining, image processing and statistical learning, we would encourage you to apply for a postgraduate research position at either MPhil or PhD levels. To be considered for these positions, please attach your CV, academic transcript, outstanding record of published work and relevant research experience (if any) to: Xiaofei Liu.