Leaders’ Insights: Alan Smeaton, Insight Centre for Data Analytics
26 October 2016
Prof Alan Smeaton is the director of the Insight Centre for Data Analytics and is a professor at the DCU School of Computing.
He leads a team of 85 people at the research centre and teaches both undergraduate and postgraduate courses at the university.
As a key figure in data science in Ireland with a variety of roles to his name, it’s hard to imagine that Smeaton “ended up in computer science almost by accident”.
Describe your role and what you do.
The great thing about my job is that I have so many roles and the diversity among these makes each day different, and interesting.
Firstly, as a professor at Dublin City University (DCU), I teach both undergraduates (first year, first semester) and postgraduates (at master’s level). I supervise student projects, I work with colleagues and I do the typical university administration. I also supervise my own research students and carry out my own research, with them.
As director of the Insight Centre for Data Analytics, a Science Foundation Ireland research centre at DCU, I oversee a team of about 85 people including PhD students and postdoctoral researchers all working in the area of data analytics; as well as an operations team, and that involves both leadership and administration.
As a board member of the Irish Research Council (IRC), I’m involved in helping to plan the strategic direction of the council as well as some other decision-making around council activities. I also like to promote the council and its role as a funding agency whenever I can, highlighting the fact that the IRC funds almost 1,400 researchers in Ireland at any given point in time.
IRC’s board meetings take place around the country and when we visit a given university or institute of technology for a meeting, we get to meet those students and other researchers funded by the council, which is great.
I’m also on the Scientific Committee of COST: European Cooperation in Science and Technology, the EU funding program with a Horizon 2020 budget of over €300m. My role on the committee is to help oversee the disbursement of this budget.
I’m an active member of the Royal Irish Academy, chairing one of the discipline committees and, finally, as an academic researcher I also do my own research work, working with colleagues abroad, travelling to present at conferences and workshops, and reviewing research papers and grant proposals from across the world.
How do you prioritise and organise your working life?
So, my life is busy, but it’s very varied, which means prioritising and organising things is really important, in order not to get snowed under.
I’m fortunate to have a great team working with me, and I can delegate and pass on things to others to help me, but for my own work I just have to multitask; put limits on the amount of time I will spend on a given task and get things done. I do feel like a circus juggler sometimes.
‘The challenge is in damping the expectation, allaying the fears and finding just the right level of expectation from what data science will bring’
What are the biggest challenges facing your sector and how are you tackling them?
Over-hype, fear and expectation management.
There’s a growing feeling that data science is the new artificial intelligence. People think of autonomous vehicles, or personalised medicine, or accurate automatic media analysis and think these will make the world a better place – but at the cost of replacing human jobs by automated processes. The challenge is in damping the expectation, allaying the fears and finding just the right level of expectation from what data science will bring.
On a personal level, my biggest challenge is in keeping up with developments and knowing what is happening and what is important in my field. A good example is in my own area of automatic description of images and videos. For years, we have been making incremental but slow progress in automatic tagging of images.
As a result of using deep learning and convolutional neural networks just a couple of years ago, we’ve made more progress in three years than in the previous 30 years and we can now automatically caption images and video, ie describe what is happening. It doesn’t always work but, when it does, it’s as good as a human can do and in my area this is a game-changer.
Keeping up with these developments is the challenge, but that is what is keeping me fresh, intellectually.
‘It is hard to think of a sector or an application that cannot be improved by appropriate use of data science techniques’
What are the key sector opportunities you’re capitalising on?
The greatest thing about data science is that it can be used in almost all sectors. In fact, it is hard to think of a sector or an application that cannot be improved by appropriate use of data science techniques.
However, the worst thing about data science is that it can be used in almost all sectors, which means people are tempted to just redeploy a technique that worked in educational analytics, to, say, agriculture or transport. That’s kind of lazy and not always the best use of the science.
I guess this comes down to us not yet having enough people skilled in the area, so new initiatives – like the degree in data science that DCU is launching – are very welcome.
What set you on the road to where you are now?
I started out wanting to do physics, then ended up in computer science almost by accident, and since then I’ve never looked back. What keeps me on that road is the support I get from family.
What was your biggest mistake and what did you learn from it?
I haven’t discovered my biggest mistake – yet!
How do you get the best out of your team?
Difficult one to answer but the most important thing is that they are happy doing their work, because if they are happy, then they will perform better – and if they are happy, I am happy.
STEM sectors receive a lot of criticism for a lack of diversity. What are your thoughts on this and what’s needed to effect change?
I’d actually be critical of those critics as I don’t think this negativity is deserving. I’ve always been of the view that the most challenging and the most impactful research is as much the margins of a discipline or the intersections between disciplines. I’m conscious that challenging and impactful don’t always go together and shouldn’t have to, as challenging can be blue skies or frontier-type work, while impact is normally associated with measuring output and KPIs, and not all research should be like that.
Within my own work, I have diversified and applied STEM knowledge in a wide range of areas. Many years ago, I worked on digitising and searching images of old Irish manuscripts. More recently, I worked on areas like digital journalism, digital technologies helping people with memory issues, digital technologies for helping elite and non-elite sports performance, digital technologies for behaviour change like diet adherence, digital technologies for energy management, for assessment of self-harm pre-disposition, for support in student education, and more.
I’m seeing more and more work which is inter- and cross-disciplinary in nature being funded by funding agencies – including our own in Ireland, like the IRC – and this is slowly effecting the change and making work like this become more like the norm.
‘The great thing about my job is that I have so many roles and the diversity among these makes each day different, and interesting’
Who is your role model and why?
A role model is somebody to follow, somebody to aspire to being or being like, but I don’t have such a hero. There are many people I admire, but nobody else I would like to be.
What books have you read that you would recommend?
I’m not a great book reader in the traditional sense but I read a lot online and, currently, I’m loving Nate Silver’s website, FiveThirtyEight and his coverage of the US presidential election. Silver is a data scientist who made his reputation from an analytical dissection of the statistics associated with US baseball, and made his name when he applied this to the Obama re-election campaign – accurately calling the result of that election before the votes were cast. He’s repeating this for the current election, assimilating all the polls and surveys as they are published into a complex prediction model for the election outcome. It makes for compelling reading – as good as any newspaper or book.
What are the essential tools and resources that get you through the working week?
Downtime. I need to be able to walk out of the office when there is a lull in the day, go somewhere else on campus and meet people not from my area for a tea break, or go for a run or a swim, and this downtime helps me recharge and get through the week.