Starting his PhD in February this year, Adam Mater is the latest RSC student to win a Westpac Future Leaders Scholarship. Adam, who is working in machine learning, decided to apply for the scholarship while studying his honours at ANU last year.
“We already have quite a few Westpac scholars so I was very exposed to the programme,” Adam explains. “Talking to others I got excited about the potential that the scholarship would offer me.”
“Even only a few months in, it’s already paying off. The money is going to partly pay for living costs and other expenses. Then I will be going to the US at the start of next year, hopefully for a month, which is partly focused on training my own skills and partly on trying to get corporate connections established with the potential for collaboration.”
Adam is Canberra born and bred, living in O’Connor for his entire life. He went through the public education system, getting interested in chemistry after being inspired by a teacher in year nine. He then decided to come to the ANU.
“I wanted to be an engineer, so my undergraduate started out as a joint engineering and chemistry degree. I didn't like it at all. Half way through first year, I dropped out of the engineering component, and then transferred up to PhB in science. “
“At the end of my second year I did my first project with my group. I started doing computational chemistry and I really liked it. I wasn’t enjoying lab work at the time, so it was a case of perfect timing I suppose.”
It is this interest in computational chemistry that has driven Adam ever since, one that has recently taken him into the direction of machine learning.
“I started my honours year with a project that revolved around traditional computational chemistry, and part way through ended up watching a video explaining machine learning. I thought that it made a lot of sense, and I could see many applications for it. I asked my supervisor if I could alter my project to focus on it, and with her approval started work on it. Fortunately it ended up going fairly well.”
“For my honours, I basically tried to give myself and everyone else in the RSC an introduction to machine learning and how it can be applied in chemistry. I did everything from the ground up, managing to get our practices close to industry standards for the systems I used.”
Adam is taking this interest into his PhD program, explaining that his project will be examining the potentials of machine learning in solving chemical problems.
“There's a few different roles for machine learning in chemistry,” Adam explains. “My honours focused on one particular application, which was developing fast molecular property prediction. I was trying to get a machine to learn how to quantitatively predict properties based off QM data sets. In doing so I managed to get close to literature standards for systems with a similar architecture. “
“The PhD is partly focused on that, but leaning more towards the intuitive and natural problem solving side of things. AI, in its more traditional sense, is more a computer that can learn to solve a problem rather than a computer that can predict a number. So I'm leaning more towards that -- trying to capture human intuition in chemistry.”
In doing so, Adam’s PhD is primarily methodological in its focus, but one that could have broad implications.
“In terms of my PhD, I'm more interested in developing a toolkit for solving broad chemical problems. Then the idea is to demonstrate this potential on specific problems that are either of interest to me, or to my group more broadly.”
“If I was going to talk which particular problems, one is retrosynthetic analysis. The common analogy is like giving someone a cake and trying to tell them what ingredients the cake started out with. That's what chemists face when they look at complex molecules and try to break them down into simple building blocks which is the reverse of a typical reaction scheme.”
“There's some recent work getting AI to solve these problems, but more can be done. Contributing to that field would be a great achievement. You could then expand this to try to explore additional complexities of chemical space, with things like narrowing down expensive conformational searching. You deal with this brutally large expansion of the number of possible orientations a molecule can adopt, what is commonly referred to as a combinatorial explosion. Most of these structures are so energetically unfavourable that they can be immediately discounted, however it is rarely apparent from just looking at the structure. So I’m also working to try to limit the number of necessary calculations using AI.”
Adam says that the Westpac Scholarship is already providing him with opportunities in this research he would never have received otherwise.
“I think the biggest thing is the network of people. I must admit it sounds clichéd, but it is certainly true. Through the scholarship, I have been able to meet other scholars from around the country. We have very different research interests but we have very similar aspirations in terms of where we want to go and how we want to benefit our fields and Australia more broadly. So having that network, and learning from all the other scholars, is really amazing.”
“Winning the scholarship was a real honour, and I’m really excited about the opportunities it is bringing.”