Introducing Mar

Hi Mar, welcome to the team! What perspective do you bring to MDAP?

Thanks, Jo! Well, I started my tertiary education in Argentina in pure mathematics and soon found my passion in applying those technical skills to real-world data. I transitioned into a PhD in sensory neuroscience where I studied how brain cells process visual information.

Then I moved to Melbourne and worked first on neuroscience education and public outreach and then on public health research. I developed a mathematical model to determine the optimal geographic allocation of medical diagnostic equipment for the World Bank, and used self-report and record linkage data from multiple databases (Medicare, PBS) to help evaluate the impact of Melbourne’s first Medically Supervised Injecting Room for the Victorian Government.

I am passionate about using, and helping others to use, computational and mathematical tools to improve the quality of research and development, support evidence-based decision-making, and automate repetitive and error-prone data processes.

What are you working on at MDAP right now?

I’m working on a collaboration using machine learning algorithms for clinical and preclinical diagnosis of neurodegenerative disease using speech data, led by Dr Benjamin Schultz from the Faculty of Medicine, Dentistry And Health Sciences.

For this, we use classification, a machine learning approach in which we teach the computer which speech recordings correspond to which diseases, and the computer then uses this learning to diagnose new speech recordings.

The unique challenge with this problem is that we have multiple speech recordings for each patient, whereas most current machine learning methods were developed for data that are mutually independent. I am now learning about approaches that consider the clustered nature of datasets in order to fully capitalise on the available information.

What are some of the solvable, difficult or wicked problems on your horizon?

I’m interested in helping to address some of the reproducibility and replicability problems in research. We need to support our collaborators with incorporating best practice techniques for scientific computing. This includes, for example, setting up pipelines, automating repetitive error-prone processes, using version control, and code review.

This is an incredibly difficult wicked problem, but I also believe it is ultimately solvable!

Can you share with us your latest adventure outside of MDAP?

Last year I was lucky enough to spend a month travelling through Vietnam and Cambodia with my partner and my mum. We started in Hanoi and slowly made our way down to the Mekong Delta and across to Siem Reap. Everywhere we went was amazing and different, as was the food, and most importantly: the coffee!

Given our global and local context, what would you most like to explore, challenge or innovate in your future work?

I envision a future in which lack of “technical” skills does not stop anyone from innovating.

I think MDAP can make a huge contribution in ensuring that The University of Melbourne is at the forefront of interdisciplinary work and I look forward to being a part of that. I also look forward to helping researchers use different tools to optimise their work — sharing knowledge is my passion.

 

If you have a research question for Mar or the MDAP team, please get in touch with us.
And keep in touch with Mar, the MDAP team, and our interdisciplinary pursuits over on Twitter.