A chemistry set from Santa catalysed this researcher’s love of science. Now, she is solving healthcare challenges with innovative machine-learning tech.
Dr Alison O’Connor is a postdoctoral researcher with the Biocomputing and Development Systems (BDS) group at the University of Limerick (UL). She also works at the Science Foundation Ireland (SFI) Lero Research Centre for Software as a co-principal investigator on a project called Alter.
O’Connor graduated with a degree in aeronautical engineering and a postgraduate diploma in advanced materials before joining the European Space Agency (ESA) on their young graduate trainee programme. Here she investigated how hydrogen environments interact with titanium alloy.
“The Ariane rockets use liquid hydrogen as a fuel source, and this is stored in titanium tanks. Titanium is reasonably lightweight and corrosion resistant, making it an attractive alloy for this type of application, but it is sensitive to hydrogen environments,” O’Connor explains.
“My research investigated how the level of hydrogen would affect the material performance.
“I absolutely loved my time in the Netherlands; my colleagues at ESA were superbly supportive of me as a young researcher.”
She then worked as a senior project manager at a UK consultancy company where she specialised in structural integrity assessments, focusing on metal fatigue.
“Metal fatigue occurs in almost all industries and so I was able to apply my knowledge across to a wide variety of challenges including offshore wind turbines, oil and gas challenges, railways, bridges and even trampolines!”
After this time in industry, O’Connor decided to fulfil her dream of completing a PhD. She studied mechanical engineering focusing on clad-welded steels for nuclear applications. She followed up this research with postdoctoral work on how crack top constraint could influence failure in metals.
During this time, she started learning Python programming language. “I was hooked almost immediately because I realised that programming elevated my ability to analyse and showcase my research work.
“I felt that machine-learning methods could better elucidate the key factors that inform material failure mechanisms.
“We could use the outputs from our machine-learning investigations to perform targeted experimentation that would push our understanding of fundamental science forward, opening even more future research avenues.”
She received a Marie Skłodowska-Curie Actions (MSCA) fellowship award to continue her research.
“During my MSCA, I quickly realised that while I was good a programming and had a strong mathematical background, the application of machine learning was more complex than I had anticipated. I didn’t quite achieve all the targets in the timeframe I had anticipated but I made sufficient progress to know that this was a viable method for achieving those targets. I knew I needed to learn more about machine learning, and I needed to learn it from experts in the field which is what led me to my current position.”
Tell us about your current research.
My current research is in the field of explainable artificial intelligence (XAI). Many of the machine-learning models that are in use today give astoundingly accurate results, but users have very little understanding of what led the model to that decision. This is problematic for research areas that require high levels of understanding, fairness and consistency such as healthcare.
I’m investigating how we can use XAI to provide additional information that helps users understand machine-learning model decision-making processes. One of the biggest applications of this technology is in Industry 4.0 applications where factories use sensors to assess things like product quality. In these cases, we want machine-learning models that tell us not only when a product we created is of bad quality, we also want the model to indicate where process adjustments can be made to improve product quality and how much that adjustment is likely to affect the final quality measurement.
The Alter project investigates how digital technologies such as machine learning and artificial intelligence can be leveraged to optimise our hospitals more effectively. One of the key interests we have is in modelling patient flow. We want to analyse how patient demand changes as a function of time so that we can ensure adequate resources are always in place. Imagine a machine learning model that assesses the number of patients today and not only extrapolates to the expected number of patients tomorrow but also provides a suggested number of doctors and nurses that can manage that level of demand effectively.
In your opinion, why is your research important?
Existing algorithms that can predict what TV shows we will enjoy or what music we should listen to next do not actively harm us – we can always turn off the TV show or skip that track, but what happens when we start to apply these algorithms to other aspects of our lives? My research is crucial because it enables human decision-making from data-driven sources.
The impact of my research is both broad and meaningful. In terms of Industry 4.0 applications, my research will help companies become more sustainable, reduce waste and increase safety. In healthcare, we can ensure that patients are seen in a timelier fashion, that adequate resources are in place, and enhance communication between hospitals and community care teams.
What inspired you to become a researcher?
I’ve always been curious about the world but one of my core childhood memories is of a chemistry set I received from Santa. I can’t be sure that I asked for it or whether Santa delivered it as a surprise, but I couldn’t have been older than 8 or 9. The set had lots of different chemicals, but my absolute favourite was the methylated spirits. I used to grow copper sulphate crystals using the methylated spirits to heat water and dissolving the copper sulphate. I loved the blue and purple colours and how the purple liquid would burn away, and the blue would dissolve in water leaving behind a gorgeous crystal. The sad thing is I don’t think they sell these kits anymore – though to be fair even as a kid I knew that it probably wasn’t all that safe to have harmful chemical that were, at the very least, flammable but honestly that just made it cooler.
What are some of the biggest challenges or misconceptions you face as a researcher in your field?
I came from an all-girls secondary school to a pretty-much all-boys undergraduate class that was really challenging from a personal development perspective. Alot of the time in undergraduate I struggled with the material; I had never encountered subjects like engineering drawing, but the others had – it felt like everyone except me ‘got it’. I didn’t want to ask for help because I was already so visible as the only girl, and I didn’t want to feed into that idea that girls couldn’t hack it. It was a very isolating experience at times and that was exacerbated by the lack of female representation at faculty level.
As a researcher I’m cognisant that things are changing (albeit slowly) but I do wonder about our female undergraduates and whether their experience differs from mine. Most undergraduates have limited contact with the research side of institutions so they’re likely unaware (just as I was) that there are many female STEM researchers around if you know where to look for them.
Do you think public engagement with science has changed in recent years?
I think that scientists have come a long way in terms of marketing and promoting their research. Social media sites particularly Twitter (now X) and Instagram have changed how researchers interact with society and audiences. I would love to say that I use social media effectively to promote my own research but that would be a lie … Personally I’m a little bit too self-conscious for the social media scene. I like to engage with people on a more personal level. I love participating in outreach events at local schools and enjoy conversations with people at events and conferences.
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.
