Image: Kevin McDonnell
Lero’s Kevin McDonnell explains how usable data is the key to doing research that can drive positive change.
With a primary degree in computer science and software engineering from Maynooth University and an MSc in artificial intelligence (AI) from University of Limerick, Kevin McDonnell was well placed to take on a PhD in AI and machine learning at the Lero Research Centre for Software and the University of Limerick.
One of the areas of his research is on the safety of electric vehicles. In a recent research paper, McDonnell and his colleagues assessed the different risks associated with electric and carbon vehicle use to provide insights and improve safety measures for car manufacturers and policymakers.
“I’ve always wanted to do something that benefited the public,” McDonnell says of his decision to engage in scientific research. Here he tells us a bit more about his work.
Tell us about your current research.
After completing my MSc in artificial intelligence, Lero and University of Limerick gave me an excellent opportunity to research and investigate machine learning models and their applications to vehicle prediction risk. It was the height of the Covid-19 pandemic, and I worked as a software engineer. Initially, I was hesitant, but taking this position was one of the best decisions I ever made.
Since 2020, my research has explored the regulations and usage of vehicle telematics data, novel machine learning methods for claim prediction, electric vehicle risks and explainability in machine learning. I have three published articles, and I’m working on my fourth paper and my Viva.
In your opinion, why is your research important?
My research touches on multiple important topics. The most significant finding of my current research is that electric vehicles are more likely to be involved in at-fault collisions than traditional vehicles. This research should assist regulators and manufacturers in reducing road-related injuries and deaths, something I’d very much like to see.
Additionally, I explore model interpretability in machine learning, promoting ethical standards and transparency for decision-making systems, areas where I hope my research will have further positive impacts.
What inspired you to become a researcher?
No one memory really stands out. I’ve always wanted to do something that benefited the public, and research or teaching seemed to be the best avenue to do that. For one reason or another, I stayed in industry far longer than I anticipated. But I’m very grateful to be in a privileged position now producing research.
What are some of the biggest challenges or misconceptions you face as a researcher in your field?
Data procurement is the hardest part of research. Even working with a private industry partner with access to millions of datapoints proved challenging.
Getting access to good and usable data is the core of influential research. Without it, it’s hard to prove our assertions and using synthetic options is always prone to bias.
Do you think public engagement with science has changed in recent years, particularly since the Covid-19 pandemic?
I started my research during Covid-19, so I’m unsure what it was like beforehand. But I think research engagement can be non-existent or all-encompassing with no scope in between. I’ve benefited from working with Lero, and they help find the best routes for public engagement. Without them, it’s certainly much harder to reach a wider audience.
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
