By Jeannine Meyer
If you are currently active in the field of artificial intelligence (AI) or have a hankering to explore and understand the depths of machine learning, then you have most probably heard of recent PhD graduate and co-founder of Brisbane’s own AI meetup, Natalie Rens. What is even more intriguing is how this South African born ballet dancer turned neuroscience extraordinaire came to be one of Australia’s most up and coming AI experts.
Upon entering university life, Natalie first saw herself as an adventurer who would take a trip to the Amazon, discover unknown exotic plants and ultimately create brand new pharmaceuticals out of them. As fate would have it, her travel log and academic journey would deviate from where it originally began upon the completion of her undergraduate in biomedical science in England. Natalie continued her studies by undertaking a Masters in neuroscience in Portugal and France, which then led her over to Australia where she continued to pursue her interest in neuroscience and thus began her PhD at the University of Queensland in Brisbane.
So, how does one write and submit a thesis in which they combine the works of both neuroimaging with machine learning? Natalie explains that she was first introduced to machine learning through one of the new techniques that had come out in neuroscience called “multi-voxel pattern analysis” (a tool Natalie used throughout her own research) which is presented in the context of “brain reading” programs that showcase how certain mental states can be picked out and translated. Natalie’s first study was created within virtual environments that revealed where in our brain the information for upcoming complex decisions was actually stored – similar to how Alice in Wonderland is required to find the right lock that fits the tiny golden key! Her study (just published – https://www.frontiersin.org/articles/10.3389/fnhum.2017.00610/full) showed widespread brain activity when individuals prepared to make their own choices, the opposite to when being told to follow instructions – (Bear with me I was just as astounded!).
Throughout Natalie’s next studies, she states she went more philosophical, saying she looked at whether you could decode whether somebody had another option available to them or not – with the idea, that if we navigate flexibly in our environment then we would store all the options available to us. Natalie found that she could decode whether an individual making a decision had alternative options available, but also, somewhat controversially, that the strength of this decoding correlated to how free they said they felt when making their decision.
After gaining some insight into Natalie’s studies, I wanted to immerse myself completely into her story. I personally find AI to be quite a specific yet complex area, and so I was curious as to what led Natalie here. She claims that while you are in academia you are in a bubble of theoretical knowledge of science, however one day she heard that a cohort from San Diego were visiting UQ and were talking about the future of emerging technologies, including AI. It was then that Natalie discovered how significant the impact would be as she came across a powerful discussion amongst the panel surrounding AI and how important it was for companies and society alike to be aware of the engineering and increase in computers that attempt to appear more human and thus, replicate human intelligence. Her fascination led her to Silicon Valley in 2016 where she became passionate about the idea that people needed to have an understanding about AI in order to not fall behind but to also be able to utilise it. As Natalie says, “it allows you to do everything, better”.
I wondered what the defining characteristics of the Australian AI market are and are there any top courses that a novice, someone like myself, could go to learn more about AI? Sadly, Natalie tells me, that in terms of AI courses here in Brisbane, it does not look too promising at the moment and that it is actually quite a problem as there are currently no tertiary institutions in Queensland that promote specialised AI programs, however Natalie recently spoke to some of the universities here who are looking at introducing programs in the near future. With the help of the established AI community, Natalie says they recently finished running the first Brisbane AI deep learning project course. In addition to developing their own projects, participants followed an online course from Fast.AI, a research institute dedicated to making deep learning more accessible, which is run by its democratic AI founders Australian data scientist Jeremy Howard and American academic Rachel Thomas. The 8-week project course just concluded, however she intends on expanding this into a series of courses this year, encouraging people to come along and learn.
Natalie explains that Australia is slightly disadvantaged in the way that “we don’t have the numbers that other countries do”. She goes on to say that, we do not have federal policy yet on AI development and technology start-ups here also struggle to find investment, in comparison to bigger hubs such as Silicon Valley. However, she says, “We can actually grow a good independent scene without having to necessarily face all the competition that they do”. She believes we have an advantage, especially in Queensland, which is home to a strong health community that boasts some of the best health research such as biofabrication research, robotics for healthcare and bionics research companies. With a confident tone, Natalie says Australia has the opportunity to just grow and start a niche start-up AI scene here – and that is exactly what Natalie has done by successfully establishing the first-ever Brisbane AI meetup.
The Brisbane AI meetup was created post attending a conference in California two years ago where Natalie asked the question, “There is this concern about the fact that if only technology companies really know what is going on and they are the only ones utilising this technology, what are we going to do?” Upon her return, Natalie attended a Brisbane robotics meetup where she spoke to Juxi, who would become future co-founder, about the idea of having a place where people could come together to learn about AI. The idea grew rapidly – to the point where they had 150 people attend their first ever meetup. The representation of females, however, within the Brisbane AI community is small, reflecting global numbers that stand at 13% – a percentage that Natalie hopes will increase as word gets out, as there is a huge demand for women in this area.
Natalie expresses that she knows there are many women out there in this environment and should you want to be involved in the community then simply reach out to Natalie or to anyone else in this space! “We are all there to support one another”. Natalie says she hopes to host another “Woman’s cocktail hour” before the next Brisbane AI meetup so that she can get more women together and form a consistent, ever-growing group of female AI leaders!
You can visit/contact Natalie Rens at the Brisbane AI website or sign up and head along to the next meetup – https://www.meetup.com/Brisbane-Artificial-Intelligence/