Adelaide-based Svelte Studios turned to Carnegie Mellon University student interns to help develop a mobile application to detect early stage mental illness.
The significance of this is that one in five Australians between 16 and 85 will experience a mental illness in any given year. Alarmingly, 54 percent of them don’t seek treatment, go undiagnosed, and fail to get help in the critical early stages.
The Svelte Studios team had built a protype called Frank, but one key element was missing: they lacked the machine learning resources and expertise.
In November 2018, armed with a proof-of-concept and an ambition to make a difference, Svelte Studios met with Carnegie Mellon University’s career services team to develop a machine learning internship.
The position was advertised to all CMU-A students seeking internships, and 6 students were shortlisted and interviewed.
“Initially we were going to take on one intern, but we were blown-away by the quality of the applicants, so we had a re-think and decided to take on two”, said Danny Connery, Svelte Studios’ Strategy Director.
“In the end, the candidates were so good we decided to take three!” he said with a laugh.
The internship focused on developing a component of the software that could interpret written text and find the emotional sentiment within it.
The three Masters of Science in Information Technology student interns, Liqian Ma, Tianyue Ma and Jiyuan Zhou were set the task of:
- researching possible machine learning options;
- understanding the methodology needed to interpret the data;
- designing, developing and deploying cloud-based machine learning programs; and
- developing the back-end infrastructure for the system.
“The internship was fantastic because we worked on a real-world project building the back-end machine learning components of the application”, said Liqian Ma.
“And 80 per cent of what we did in the internship, we learned in our Cloud Computing course – so it was gratifying to be able to apply our classroom learnings in the real-world.
“We weren’t looking to undo the good work Svelte Studios had already done, but our preliminary research pointed to using cloud computing and processing the data on a server instead of on a device”, said Liqian.
Thanks to the CMU-A interns, Svelte Studios is now well advanced in the development of the important first stage of the mental and emotional health detection application.
“Having the interns was amazing. We always expected a good outcome, but their contribution exceeded our expectations!”
“They all worked extremely hard, and having the education, experience and confidence to actively contribute right from the start - making recommendations and giving us plenty to think about.
“They were really confident in the way they wanted to approach things, and that made us change direction. It worked out really well. Because of their confidence, and competence, we have been able to accelerate the project”, said Danny.
What’s next in the development of Frank and the relationship with CMU-A?
Danny says, “Well, our relationship with CMU-A continues on strongly. We were able to keep Liqian and Tianyue on in a part-time capacity beyond their internship, and we have now taken on another two interns to develop the next stage of the application”.