World class machine-learning helps solve industry problems

That’s where Professor Damminda Alahakoon comes in. Damminda is the director of the Research Centre for Data Analytics and Cognition at La Trobe University, and a big believer in his team’s unique value to industry.

“We’re based in the business school, but we do research across the university, and nationally as well as internationally,” he explains. “Most of us are computer scientists and engineers, so a very highly technically trained group of people.”

Damminda and his team of PhD candidates work on cutting-edge research in AI and machine-learning. He describes his research as fundamental: it uses conceptual, theoretical and even philosophical modes of thinking to find real-life solutions for industry-level problems.

Solving problems of all sizes

Among the industry partners of Damminda’s centre are corporate giants like AGL, medium enterprises like Melbourne-based uniform manufacturers Shiny Embroidery, and brand new start-ups that are just finding their feet.

For AGL, Damminda’s team used AI to optimise the energy company’s consumer profiling, to assist with costing, billing and strategic planning. To do this, the team used machine learning-based profile development to innovate sophisticated, detailed and granular profiles for AGL’s customers.

“For AGL, we had more than 1 billion data points using our algorithm,” he says. “We used completely new algorithms that were developed in the research centre, but delivered an industrial level solution.”

For apparel company Shiny Embroidery, a government innovation grant funded Damminda’s team to solve some very practical problems.

“They have large five warehouses in Fairfield. They wanted a better, smart phone-based order picking system. So we built it for them.”

Damminda says for start-ups in particular, there is one enormous upside to using a research centre instead of trying to hire the right person.

“It's very hard, especially with this very specialised expertise, for a start-up to hire. Their ability to get the right people, or number of people, is limited, and they probably will not need these people for the whole year.”

“We have the expertise, with very strong industry skills. We can start a project with about 10 or 12 students, and through their PhDs we explore some very innovative ideas.”

The collaboration with Shiny Embroidery came about because of connection with North Link, who are co-located at La Trobe University, in Bundoora.

“There is a potential to work with the range of different companies,” Damminda says, adding that word of mouth and personal connections are the best ways to find the right partner.

Human behaviour trials on the cards for La Trobe

In terms of collaboration, Damminda is excited about La Trobe’s plan to transform the Bundoora campus into a thriving, dynamic University City of the Future. His team’s role in this concept is already established, contributing to the university’s aim for zero carbon emissions by 2029   .

“We are building a lateral energy analytics platform, called LEAP, for monitoring the current energy usage across the university campuses,” he explains. “We are planning to expand into gas usage, water, and, when we have the sufficient sensors, crowd movements and other activities in the university.”

The sensors to monitor human behaviour are, for Damminda, the most exciting prospect the University City of the Future holds.

“The University City of the Future brings with it this idea of living lab. It's a beautiful concept because we researchers can build things and, before we take it out into the world, we get the opportunity to trial within the university itself.”

“That is the most exciting thing for me. In censored environments, they're collecting data and human behaviour to such a detailed level of granularity; it's an amazing opportunity for me and my team.”

Damminda believes in the concept wholeheartedly, adding that La Trobe is in a unique position to become the first Australian university to create this campus-to-city transformation.