The overarching goal of my research is to build technologies that enhance people’s cognitive abilities and help them navigate a world full of information and distractions through pervasive and ambient computing.
In my research, I apply user-centered design principles to design, implement, and deploy ubiquitous computing systems, specifically in the following application areas:
Cognition-aware systems are technologies that sense users’ cognitive states, model systematic variations, and adaptate their interface and behaviour. Using biophysical sensors and interaction data, this work focuses on quantifying and modelling users’ attention, alertness, and general cognitive performance. By considering the cognitive state as a context variable, we can build better reading interfaces and learning systems while preventing information overload. In recent years, however, it has become clear that optimising the information bandwidth between human and computer is also about assuring the quality of information processing. Together with behavioural psychologists and the information retrieval community, this work has pioneered a series of workshops at CHI about technology design that supports critical thinking and the role that cognitive biases play in system interactions and information consumption. The resulting bias-aware systems are capable of detecting and helping to mitigate the effects of cognitive biases in opinion formation and decision-making. This work has recently received an Honorable Mention Award at CHI and has led to the organisation of two Dagstuhl Seminars to formulate and advance a research agenda around cognitive biases and potential vulnerabilities in the context of misinformation and social engineering attacks. This work was funded by an Early Career Research Grant from the University of Melbourne ($20k USD).
Reading takes increasingly place on digital devices where medium, interface, and content presentation are highly dynamic. Additionally, the information age provides us with the challenges of information overload leading to the creation of new reading behaviours. This work strand investigates these behaviours and designs interfaces with the goal of providing better readability and instilling better reading habits in users. This work is mainly funded by Industry partner Adobe Research and their Documents Intelligence Lab ($28k USD) and has resulted in publications at CHI, IEEE VR, and MUM. The work takes a holistic approach to reading interfaces, i.e., on Desktop and mobile devices but also on emerging technology platforms, including virtual reality and head-mounted displays, and custom-made hardware. Eye movement characteristics are used to infer reading depth, engagement, and mind wandering, while large language models automatically create previews and reviews to prime readers for better text comprehension and help them recover from reading interruptions. Together with Adobe, this work also contributes to a larger, consorted effort of researchers from both academia and Industry to research digital readability and advocate text parameter customisation (see The Readability Consortium). People’s reading behaviour is changing due to digital reading devices and new media formats. This research contributes to the understanding of how this behaviour changes through the development of novel measures and the investigation and design of interfaces that prioritise information gain over attention capture.
Technological developments in ubiquitous sensing, data source integration, and machine learning offer new opportunities for patient care. Technologies provide the means to continuously measure individuals’ behaviours and health states to support the achievement of behavioural goals. This research strand, currently co-funded by an NHMRC Ideas Grant ($800k USD), focuses on the development of monitoring technologies and conversational interfaces that help instil and sustain healthy behaviours. Novel sensors are developed using thermal cameras to monitor hand hygiene quality and near-infrared spectroscopy to detect pill and infusion content. By devising a trigger mechanism using inaudible audio to trigger Google Home apps, the off-the-shelf voice assistant becomes proactive. Additional sensors are used to detect opportune moments to engage patients in their homes based on their proximity, vitals, and privacy settings and are used to collect patients’ self-assessments throughout the day. Smart monitoring paired with a direct intervention line to caretakers can help patients manage their chronic diseases and keep them out of the hospital longer. We recently established a joint PhD graduate program between the Berlin University Alliance and the University of Melbourne with $205k USD in funding for a collaboration between the computer and behaviour science.
I am always looking for ambitious PhD students interested in any of these areas. If they strike your interest and you are excited about working in Australia and/or Germany, do reach out!
Current Project Funding
|NHMRC Ideas Grant||Harnessing information technology to improve self-management behaviours and health outcome in people with heart failure: A smarthome ecosystem Living Lab Study||1,239,220||2021-2025|
|Melbourne-Berlin University Alliance||Digital Technologies to Measure and Promote Sustainable Health Behaviour Change||310,000||since 2020|
|ADOBE Documents Intelligence Lab||Quantifying Reading Behaviours on Mobile Devices||42,000||since 2020|
Past Project Funding
|MSE Platform Interdisciplinary Grant||Detecting Cognitive Biases with Biophysical Sensors to Battle Misinformation||30,000||2020|
|DAAD (German Academic Exchange Service)||Adaptive Learning Interfaces||24,800||2019|
|UniMelb Automotive Engineering Graduate Program (AEGP)||From Future Fuels to the Future of Transport||644,000||2019|
|Early Career Researcher Grant (ECR) Scheme of The University of Melbourne||Critical media - Building Technologies for Media Literacy and Depolarisation||38,000||2019|