How two students found machine learning PhD opportunities in Australia
After undergraduate studies in Russia and a double masters’ degree at LUT University (Finland), Violetta Shevchenko knew two things: she wanted to obtain a PhD in machine learning, and to live in Adelaide.
She started looking for opportunities.
“I had visited family in Adelaide before, and I really liked the city,” she explains.
“So I just googled Adelaide + computer + vision.”
She found a profile of Professor Anton van den Hengel — Director at Australian Institute for Machine Learning — and sent him an email asking to meet when she was next in the city.
The rest, as they say, is history. Violetta is now in her second year of a PhD in visual understanding and reasoning in machine learning. Her supervisors are Anton, Dr Anthony Dick and Dr Damian Teney.
“Violetta’s CV and transcripts showed that she had done quite well in some relevant degrees at good universities, but also that she had done a practical project that had a real outcome,” says Anton.
“The combination of good theoretical skills and strong practical problem solving is certainly something I look for in students.”
Teaching AI to explain a photo
A typical day in Violetta’s PhD life consists of a lot of critical reading of other people’s research, then designing and running her own experiments with different models of machine learning.
She is based in the Australian Institute for Machine Learning building, in the centre of Adelaide.
“My main focus is to test if AI can actually understand something about visual input,” she says.
“In particular, I’m exploring whether AI can process visual information — for example, from a photograph — and answer related natural language questions.”
The end result could be something like a tool that creates a caption from a photo.
“So maybe a person with a vision impairment would be able to go to a shop, take a photo of an item of clothing, and the AI could tell them the colour, the shape and other features of product,” Violetta says.
After she graduates, Violetta is interested in staying in Adelaide and working at the intersection of research and industry.
“I want to do something practical, perhaps in computer engineering,” she says.
Finding a high calibre supervisor
Dr Rafa Felix also studied for a PhD at Australian Institute of Machine Learning. He is now a postdoctoral researcher in artificial intelligence at the institute.
“My PhD focused on zero shot learning,” Rafa explains.
“This is a kind of AI where we try and teach a machine to be able to identify something it has never seen before.”
In seeking out a PhD opportunity, Rafa’s main criteria was to find a high calibre supervisor. Originally from Brazil, he had previously worked in industry and completed undergraduate and master’s degrees in computer science and machine learning.
“I started looking around the world, and found Professor Gustavo Carneiro at the Australian Institute for Machine Learning,” he says.
“My approach was to send him an email telling him I was interested in studying with him. But also I was very deliberate in convincing him I was a strong PhD candidate.”
Rafa included a detailed CV in his email, a compelling cover letter, and some ideas about a possible research project. He secured an Australian Research Council scholarship to support his candidacy.
Gustavo says a track record of scholarly excellence is the key factor when he looks at potential PhD students.
“I look to see if they’re applying from good schools, if they have good marks, and even whether they have already published something academic,” Gustavo explains.
“Having publications is becoming more and more important for potential PhD students these days.”
Rafa’s PhD was co-supervised by Gustavo, Dr. Michele Sasdelli, and Professor Ian Reid, and attracted a University of Adelaide Dean’s commendation when he graduated in 2019.
For now, he’s staying firmly put in Adelaide.
“It’s an amazing place for bike rides, and it’s very connected to nature,” Rafa says.
“But also it’s a very vivid city — there are parties, and you can find interesting places to go.”
Rafa’s postdoctoral position allows him to do a mix of academic and industry work at the Australian Institute of Machine Learning.
“I love it. I found myself here,” he says.
