AI, ethics and happy pets: A year of research in review

Our dependence on technology reached a whole new level in 2020. As the world went into lockdown, we went online to keep working, socialising, schooling and everything else in between.

As we spent more time than ever on devices, Faculty of Engineering and Information Technology researchers were at the forefront of computing research. They were devising new ways to use powerful technologies like AI, but also considering how we integrate technology safely and ethically into our everyday lives. Here are some of our highlights.

Ever wondered what your pet is thinking?

With owners home 24/7 last year, many of us thought our pets were living their best lives. But were they really? The Happy Pets app, which was launched in mid-2020, gave us a few clues.

Dog sitting in grass with tongue poking out

Developed in the Melbourne eResearch group, the popular app uses AI to assess how your pet is feeling, using a percentage for each of the five most common animal emotions (happy, angry, neutral, sad and scared). It also tells you their most likely breed.

COVID-19: A challenge and an opportunity for public transport

Last year our living laboratory for safer, cleaner and more sustainable transport, AIMES, continued its sector-leading work on the streets of Melbourne.

Blue string connected by nails on white surface

Arguably, the team’s challenges were greater than ever, with COVID-19 causing major changes in how we move around the city. Even after lockdown, we are using more private vehicles and less public transport than we used to.

However, in the long term we need to see the opposite trend if we are going to live more sustainably – fewer private car trips and more trips on public transport. How can we move towards this while still guarding public health? Dr Peter Sweatman and Professor Majid Sarvi argued that innovations like sensing technologies to manage the flow of passengers and AI to predict traffic activity levels will help.

Scaling up quantum computing – thanks to machine learning

Quantum computing is coming and the race is on to build the world’s first universal quantum computer. Key to this is maximising qubit efficiency, which often depends on ultra-precise placement of atoms at an incredibly tiny scale. A challenging task indeed.

A map of electron wave function patterns, where the symmetry, brightness and size of features is directly related to the position of a phosphorus atom in silicon lattice

Dr Muhammend Usman was part of a team that devised a machine learning algorithm to processes qubit measurement data with high-throughput, high precision, and minimal human interaction, bringing us one step closer to a quantum computing future.

Why we let ourselves be monitored – and how to keep your data safe

Last year saw the launch of the Centre for AI and Digital Ethics, which brings together experts from across the University to tackle ethical, regulatory and legal issues in AI and digital technologies. Our own Professor Tim Miller is co-director, along with Professor Jeannie Paterson from the Melbourne Law School.

Hand holding smartphone with binary code projected over it

Some of the centre’s experts got together to offer timely practical advice on how to keep your data private, even (or especially) when technology is inescapable, tackiling everything from the use of technology for public health purposes to the dangers of relying on algorithms to predict school scores.

Disarmament in the age of AI

Senior Research Fellow in Digital Ethics, Dr Kobi Leins, argued that the international community needs to address dangers posed by the potential weaponisation of new technologies (like AI) now. Waiting too long to negotiate norms and laws will mean the disarmament community could be caught on the back foot, finding itself responding to a crisis rather than pre-emptively managing a threat.

This is part two of a series looking at some of the most promising innovations from the Faculty of Engineering and Information Technology in 2020.

Read more about our work in AI and data.