I am currently in Zurich Switzerland, working at Google as a Researcher where I investigate novel approaches to machine learning and perception. Previously, I was a Postdoc at the Applied Artificial Intelligence Lab at the University of Oxford (formally part of the Mobile Robotics Group) working with Ingmar Posner and Paul Newman. There, the scope of my research covered various domains including multi-task learning, unsupervised domain adaptation, visual attention, model introspection and interpretability. I completed my PhD research at the Queensland University of Technology (Australia) alongside the ARC Centre of Excellence for Robotic Vision. My PhD topic was focused on the automatic detection and tracking of moving objects from video data with applications towards field robotics. Prior to my PhD I worked a number of years as an automation engineer, project leader and research associate at CRCMining and the University of Queensland. In 2008, I completed a bachelor of mechatronic engineering at the University of Sydney while working part-time as a software developer and consultant.
My general scientific interests lie in the fields of machine learning, computer vision and robotics. Specifically, I'm interested in data efficient machine learning through the exploitation of innate structured priors. Aspects and approaches include:
- Auxiliary and adversarial losses in a multi-task learning context.
- Transfer learning, domain adaptation and knowledge distillation. Particularly in relation to bridging the sim-to-real gap.
- Attention mechanisms.
- Representation learning through unsupervised clustering, autoencoding and metric learning.
- Semi-supervised learning with spatial-temporal models for problems such as multiple object tracking.