Gráinne and Louise talk to Professor Madeleine Lowery about her research at the frontier of the effort to combat devastating diseases like Huntington's and Parkinson's; Professor Alan Smeaton talks Zoom fatigue and Professor Barry O'Sullivan discusses policy influence as researcher and his plans for Insight 2.
Gráinne and Louise talk to Professor Madeleine Lowery about her research at the frontier of the effort to combat devastating diseases like Huntington's and Parkinson's; Professor Alan Smeaton talks Zoom fatigue and Professor Barry O'Sullivan discusses policy influence as researcher and his plans for Insight 2.
Louise Holden 0:06
Hello, and welcome to episode two of The Insight Podcast. Today's show is available across all podcast platforms, and if you like what you hear, we would appreciate it if you would rate review and subscribe. This great show lined up for you today, don't we Gráinne
Gráinne Faller 0:20
We do indeed Louise. We are going to be talking to Professor Alan Smeaton all about Zoom fatigue and some interesting research going on there. We're also going to be talking to Professor Barry O'Sullivan down in Insight at UCC about some of his work in AI and also how such a well travelled person is coping with lockdown,
Louise Holden 0:38
We have a full hand of professors today. We're going to be talking first to Professor Madeleine Lowery. Madeleine Lowery has been using engineering to understand the human nervous system. Her work as part of the effort to understand and treat devastating conditions such as Huntington's and Parkinson's disease.
Gráinne Faller 0:52
Yeah, this is fantastic hope you enjoy.
Madeleine Lowery 0:55
So we're interested in understanding how the body moves basically. So we're interested in understanding how the nervous system controls muscles, and then how that changes under different conditions. So for example, as we age or as muscles fatigue or also the changes that take place in different types of neurodegenerative or neurological conditions, for example, a Parkinson's disease or Huntington's disease, or maybe after a stroke. So really signals are communicated around the body using electrical impulses, so very small, little current and voltage pulses. So in many ways you could look at it like an engineering system, and you can use engineering methods to try to understand how that communication works. So we on the engineering side, the two main tools we use are, one is computational modeling. So we can build up quite detailed models of networks of neurons within the brain, or of all of the different muscle fibers within the muscle and the electrical fields that they generate, and we can use those to try to understand the mechanisms by which it works. And then by which it changes under these different conditions. And the other tool that we use is signal processing or signal analysis. So we can record signals from the body, you can record the electrical signals the muscles generate, or you can record the electrical signals that are generated by the brain, either with electrodes on the scalp or with electrodes implanted deep within it. And then we can use signal processing techniques in combination with an understanding of how the system works to try to extract information about that, so to to unpick it, if you like and get a window into the nervous system and into the neuromuscular system.
Gráinne Faller 2:26
Is this all just is it understanding how things work? Or is it understanding what is happening when things break down?
Madeleine Lowery 2:35
Both. So in order to be able to understand and and ultimately what we want to be able to do then is come up with therapeutic interventions or therapies or technologies, such as medical devices that can help then fix things, or improve function when they do break down. So the first part of it, I suppose, is to understand really how things work in the, let's say, normal conditions or under healthy conditions, then we understand what they should look like, for example, what those firing patterns of those nerves are, then that allows us to look at them in pathological conditions or disease conditions, and you can see what those differences are. And then that could shed some light on what's going wrong, and what you need to do in order to try to restore the function that's been lost, you know, be that movement and the sort of stuff that we're interested in. And then the last piece of the jigsaw is how you can change things. So what sort of an intervention you can do in order to change things. So that could be something like a therapy such as exercise therapy. So we've had some work recently working with colleagues in the Royal Hospital in Donnybrook, where they're doing specific targeted exercise interventions in people with Parkinson's disease. And we can use sensors on the muscles and on the limbs in order to track the changes that take place over time. And then we can see, does this work? And if it does, what does it affect? For example? Does it improve their gait cycle? If it does, how? Does it change the coordination of the muscles? And does it change, fine motor control, or just big movements? And then that provides some quantitative information on that. And then you can also look at different medical devices. So one that we're particularly interested in as electronic engineers is the whole area of neuromodulation and electrical stimulation. So you can electrically stimulate nerves and muscles by applying small currents or voltages to them using electrodes, either the skin surface or implanted within the body. And one area there that we're particularly interested in is deep brain stimulation for Parkinson's disease. So there they implant electrodes within the brain, and they stimulate neurons in the surrounding area, and they change the patterns of activity, and it's very effective in restoring lost movement, lost motor function, but the mechanisms aren't fully understood and from an engineering control system it's relatively simple. So you go in, you stimulate, and you use trial and error to try to set what the stimulation parameters are. But what we'd like to be able to do, is do that in a more and I suppose a more intelligent way, from an engineering system perspective. So you can record signals back from the body, understand what's going on, and then change your stimulation in in order to adapt to that.
Louise Holden 5:02
It's it's an amazing crossover of fields. You've been involved, haven't you Madeleine, in the development of a device for, isn't it to do with movement in the hands?
Madeleine Lowery 5:11
So we have yeah, that's right. So this came out of the exercise study that I mentioned. So it's a very simple device really, developed with two of my my PhD students, Ben O'Callaghan and Matthew Flood at the time. So it just uses accelerometers placed on the hand and it can measure finger contact times during the sort of tapping tests that they do within the clinic. So normally, the clinician will look at the patient and they'll judge, you know, they judge their movement in a kind of a qualitative way, a subjective way. So what we want it to be able to do is to track that over time, and to look at the changes over time. So it's a simple device that can track the amount of tremor in the limb. And also then the fine motor control and give you some hard sort of information or numbers on that, that you can use in order to quantify improvements in motor control.
Gráinne Faller 5:57
The Deep Brain Stimulation aspect, I do remember, I think it might have been Insight's, first SFI review, where you showed a video of a patient with Parkinson's who was walking down the hall with huge tremor. And then after implantation of a device, I think it was almost like a transformation.
Madeleine Lowery 6:20
That's right.
Gráinne Faller 6:20
I mean, it looked like something from science fiction, almost. What stage is that sort of technology at now is that actually in use?
Madeleine Lowery 6:28
Oh, it is. It's very much a mainstream technology. So it's been applied in about, the numbers aren't exact, but somewhere between 250, you know, somewhere around 250,000 patients worldwide. So it's an established therapy for Parkinson's Disease at the moment, and also for essential tremor. And then it's used, and also for Obsessive Compulsive Disorder, and for specific motor disorders as well, including dystonia sometimes. So it's it's very well, tends to be used now in patients who've stopped responding to medication, so patients who have responded to medication but for whom the medication is no longer effective, and then they will generally be, you know, considered for deep brain stimulation. So it's, I mean, the DBS devices, by one of the companies who manufacture them, Boston Scientific, are all made in Ireland as well. They're manufactured in Clonmel. But it's very much, essentially it's like pacemaking for the brain. Now, we're familiar with cardiac pacemaking, that's been around a long time. And it's doing something very similar to the neurons in the brain instead of the cardiac cells in the heart.
Louise Holden 7:29
That's extraordinary. And are you networked into a global research infrastructure on this? Are there a lot of people working on this all over the world?
Madeleine Lowery 7:37
There are. There's different groups working on us all over the world. We've collaborators in Europe, we've collaborators in Paris. We've different groups that we're working with in a different area. In Huntington's disease, we're working with a group, who are looking at wearable sensing, to measure physical activity, sleep and nutrition, in individuals with Huntington's disease at different sites throughout Europe. And that's been led by Cardiff University. So it's certainly an area that there's a lot of work being done in worldwide.
Gráinne Faller 8:14
Can I ask, this might seem like a quite basic question. But when somebody goes into engineering, where you are, I don't think many people it would even occur to people that the kind of research you're doing, would have, it's not a natural connection that people make, I guess, with engineering, people might think building bridges or something. But when you started out in the very early days, I mean, did you have any notion of where you would end up?
Madeleine Lowery 8:41
I was, always interested in the biomedical engineering, but at the time, it wasn't well established. So I went into engineering because I liked the the maths and that side of it. And then I went into electronic engineering from that. But I always had an interest in biomedical, I always was interested intersection between medicine and engineering. But nowadays, we have a biomedical engineering program here in UCD. And you know, many of the universities in Ireland and internationally do as well. So it's more visible to our students coming in. But in a way, the sort of when I started, your first question was kind of around the, maybe the tools that we use as engineers, so I use computation modeling, signal processing, data analysis, and that's exactly the same as my colleagues next door might be doing working on problems and on the power system or on communications. So you're applying it to a different type of system and the problems are different, obviously, and you have to change it and tailor it and you're dealing with different frequencies and things behave differently. But it's the same tools if you like that you're using and we work very closely as well with colleagues in science, in medicine, and with with clinical collaborators, of course, you know, to be able to do any of that work with patients.
Gráinne Faller 9:50
What are you.. What does the next couple of years hold for you? What are you working on at the moment and where are you going?
Madeleine Lowery 9:56
So at the moment we're working in the, well the study I mentioned on Huntington's Disease is one that we're sort of in the middle of at the moment. We have, we're still working in the area of deep brain stimulation on developing strategies. So, and algorithms that would allow you to sense information from the body, and then to change the maybe the frequency, or the amplitude,or the pulse duration of the stimulation in response to that. So we'd respond to changes in the patient's symptoms, you know, that occur within the day or over, over weeks or months. And then we also have studies started as well in trialling those, so we develop them and test them on our computational models, and then we're starting - we've small preclinical tests started to be able to test them, their efficacy, or show, proof of concept.
Gráinne Faller 10:40
It must be very exciting.
Madeleine Lowery 10:41
It is very exciting. It's very interesting. It's always interesting. Yeah.
Louise Holden 10:44
I can imagine too, that it must be very exciting to know that some of this research is being translated here in Ireland. It's not just it's not, you know, the way sometimes you can do a little piece of it, and then it tends to leave Ireland, whereas in the device development sector, Ireland is actually quite front and centre. Am I right?
Madeleine Lowery 10:46
That's right. It is it's very much so. So the medical device industry in Ireland is, is huge. It's a huge employer. And there's a lot of R&D now taking place in Ireland as well. And it's great to see our graduates coming through as well. You know, they get placements in the so our biomedical engineering students are coming through and they're getting placements in industry, and they're going off to work in industry in the area, you know, which they've trained and which they've gone into, which is really nice to see.
Gráinne Faller 11:25
And that was Professor Madeline Lowry, from insight at UCD. What an amazing interview. What incredible research.
Unknown Speaker 11:37
Isn't it really, really something the magic that can happen when you take two very disparate disciplines and work in the middle area between them really can't wait to see what happens next.
Gráinne Faller 11:45
It's really fantastic me neither. Now we're going to move on to Professor Alan Smeaton.
Louise Holden 11:51
It's been a funny old year. We have some new phrases knocking around things we never thought we'd hear. Pandemic payment, anti-mask march wouldn't have made any sense in January. Zoom fatigue is another. You've been doing some research on this in DCU Professor Alan Smeaton. Tell us what you've learned about Zoom fatigue.
Alan Smeaton 12:09
Yeah, it's prevalent, and it's everywhere. And when on the 12th of March, we jumped into lockdown, the universities right across the world, we rapidly moved to online lecturing. The default was that lectures would continue as per the timetable online on Zoom classes. And I, like many others, found myself lecturing to large numbers of students, especially at undergraduate level - at one point, I had something like 93 students on the zoom call for a lecture - and it was just me in my office, effectively talking to myself, because every one of them had their cameras switched off. And some of them, I found out later, weren't even there. They were just, there to be present to be there, but they were just totally non engaged. And this is the phenomenon of Zoom fatigue, which we get in long meetings. But we also get especially for presented content with lectures. One of the things we found out afterwards, after we had studdled our way through the through the semester, was the feedback from students is that their level of engagement with Zoom lectures, is really, really low. When you're facing a class of students, you can look at their body language, you can look at their faces. Even when you're on a small zoom call with two or three people you can look at their faces and and we're socially obliged to, to engage with whoever speaking to us by nodding occasionally, as you're both doing there, and and that sort of thing. And that works fine face to face, or even works fine on Zoom with small numbers. But when you're dealing with large numbers, there's just no engagement whatsoever.
Gráinne Faller 13:46
So what do you do about that?
Alan Smeaton 13:48
Well, you could try to make the classes smaller, or you can try and make them interesting. And what we've done, what we've learned, as academics for delivering content is that that mode of trying to force an interaction with a large number of students online, it doesn't work. So going forward for next semester, for next year and onwards, we're going to cut up our material into smaller, bite sized chunks, shorter versions of lectures, make them online and available for students to download and play in their own time. And then engage with them in dialogue in sort of reflection sessions, assuming that they've played through the lecture content. But for face to face stuff, that's where zoom fatigue still remains. We've removed the biggest instance of it, which is online lectures, but we still face it with with smaller groups and trying to engage with people and making sure that they engage with the content.
Gráinne Faller 14:40
I seem to remember Alan, when we spoke very early on when we started with Insight, you were doing work on, or people you were working with were doing work on analyzing people's engagement via sort of cameras, I think it was for customer services at the time. Is there any potential for that kind of thing for this?
Alan Smeaton 15:00
We have something in the pipeline, which is looking at exactly at that. So working with a couple of others from Insight, so working with Mingming Liu and Hyowon Lee, we're developing a system which uses the webcam on your laptop, which you currently use in zoom calls. But what we do is we capture the students face during an online lecture, a real live lecture or a live session, we capture the students face and we measure the level of facial engagement, how interested and how keen they are on the topic. And there's lots of ways you can do that. The simplest way that we've found and the way that works is to simply measure the aspect ratio in the eye. So if you draw a horizontal line from the inside of the eye to the outside of the eye, and you draw a vertical line from the top of the eye to the bottom of the eye, in other words, how open are the eyes, and what we do is, is that we measure that for both eyes in real time. And as the head is moving from side to side, as you're nodding, it tracks this. So we've developed an application which which monitors this during the lecture that a student may be taking live. And this attention level is then fed back in real time to a cloud service. And what we do is we gather this for me for you for every one of the other students who are using this system. And at the end of the lecture, what happens is we slice it all and dice it all and align it. And we're able to give feedback to the lecturer on what were the attention levels of the lecture, that she or he just gave. So you get the parts where everybody was really interested in reverted to that part, because everybody was really paying attention. And then the other parts, which were the flatlines went there was nobody interested at all.
Louise Holden 16:43
So this is a this is an aggregate This is aggregate feedback, you're not getting individuals. So you know, Jimmy was not paying attention. So we get, we'll give him a little shock. (laughs)
Alan Smeaton 16:54
Yeah, so so the great benefit for the lecturer and this, at the moment is done. It's not done live, it's done post event. So I mean, it would be great, and ultimately, it would be great to do this live. I mean, imagine giving a lecture to 100 students and seeing a little icon at the bottom, which is a slider indicating how engaged students were, that would just be it'd be a substitute, but at least be something as a substitute for, for looking at the engagement. And the other nice thing about this, the system we're developing is this, that on the zoom platform lectures are normally recorded. So you can download the the video of the recorded lecture afterwards. And what we do is we allow students to download the recording of the lecture that they sat through, which they will use for revision purposes. And then what we do is we combine their individual attention during that lecture, to just get a summary of the parts that they missed.
Louise Holden 17:48
For them an individualized report.
Alan Smeaton 17:51
Yes
Gráinne Faller 17:52
That's clever.
Alan Smeaton 17:52
Yes. So that's the real benefit. So the previous benefit is the benefit for the lecturer. So you know, she or he sees what parts were good or bad, and she or he can improve on that. But individually, a student can then get their own personalized summary. So if you sat through a one hour lecture, and you just went Rip Van Winkle for a part of it, okay, or you had a mind wandering, or there was an Amazon delivery at the door, then what you get is the summary of the parts where you were flatlining, and most importantly, where you were flatlining and others were not. So if you flatline during a part where everybody was flatline, then you didn't miss much compared to others. But if you flatlined compared to others, then you're really missed a part that others caught. And that's the summary that we give back to you of that lecture.
Louise Holden 18:39
And do you anticipate that this will have a behavioural effect? If people know that they're being monitored for their attention level? Are they more likely to pay better attention?
Alan Smeaton 18:48
That's the big question that people ask about this. And you know, that camera is on, the little green night is switched on, the LED and switch on it knows it's looking. But what we do in the system is is that we monitor the face, we calculate the attention level, and then we feed back the attention level, so we don't record the face at all. So there's no recording of the face, the only recording is the recording of the attention level. And then when that comes back to our system, it's anonymized when it's aggregated and fed back to the lecture. So the the lecturer never sees that it Jimmy or Sarah wasn't listening during that part. And then when we generate the summary, we of course, we of course, have to know who it is for, but we don't do anything with that data. Because there's nothing we can do with that data. We're just selecting segments of the overall lecture, which are highlights or lowlights in that case.
Gráinne Faller 19:39
And is this easy to sell to students? In the sense of getting people to sign up because immediately, I mean, when I heard about what you were doing, I thought, oh, my god, as an undergrad, I never would have wanted anybody to know my attention levels in lectures because it wasn't very interested in a lot of the ones that I did as an undergrad. But the fact that like, the fact that you're explaining what you're explaining is actually a great benefit. And but I mean, you know, you would have had me, I would have been out the door at 'camera,' you know if you'd tried to explain that to me at that point. Are you finding it an easy salary difficult sell to students?
Alan Smeaton 20:15
I think that there's lots of technology challenges with developing this. There's lots of meaty data analytics, problems. But the big thing would be that first 10 seconds in which we sell this idea, and if we can do this, in a way that allows them to see the benefit of it before they go out the door, and switch off, then then this, this could work. I mean, I've been thinking about ways to do this. And I'm thinking of sort of animated cartoons short, you know, a 20 second, or 30 second animated cartoon to show you what you can get from this. And it's all about you. And and, and that's going to be critical. If we can do this, and do it in a way that gets some small portion of them thinking positively about it, then that feedback that they have from that, then then it goes from person to person to person, because there's no sign up at the beginning of the semester, and then you're in. And if you miss that boat, you're out because this system would run and allow people to opt in or opt out at any stage.
Louise Holden 21:16
Very interesting. I might not be first in line to volunteer for that, but I imagine Alan's students would be more intrepid. Next up, we're going to talk to Professor Barry O'Sullivan. He can be a hard man to catch, because he's always on the road and traveling the world influencing AI research and policy wherever he goes. But he's been grounded for the last year, and we wanted to catch up with them and find out what he's been doing from his home in Cork.
Gráinne Faller 21:37
It must be a bit of a change being tied to a place for six months.
Barry O'Sullivan 21:41
Yeah, like I think I was probably traveling away maybe 40% of the time. So almost two days of every week, I would be somewhere. And you know, now in the last six months, I think I've been I think I've been out of the county once, took my mother on a spin. And apart from that, I've just been to the local town. I haven't missed it, actually. I don't see myself going back and traveling as much as I used to. So it's been a sort of a great recalibration, but at the same time, it's sort of too cognitively weird, to be enjoyable, you know
Louise Holden 22:13
yeah, yeah, there's a mild buzz of anxiety in the background. Tell me this, in your work you have just come to the end of your presidency of the European AI Association and a broader involvement with that organization too. Looking back over those two years, what have you taken away from that experience?
Barry O'Sullivan 22:30
Yeah well it's been interesting. So I was the president of EurAI, the European AI Association for two years. It finished at the beginning of this month. So EurAI is a is a large organization does about four and a half thousand members across about 30 countries. It's an association of the National European AI Associations. So it's been a great learning experience, I guess it gives you a platform that you wouldn't have otherwise. So one of the things that I tried to do in my time at your AI was get more involved in European issues and guiding policy and so on, and we achieved that. So that's been great,
Gráinne Faller 23:06
Is there a constant push and pull when you're dealing with something like AI, as a European or North American level, where you have, on one hand, people who are very keen to use the technology and implement it. You know, obviously, there are huge advances that can be made. But increasingly, the more we're finding out about it, the more we're thinking about it, the more we're conducting research into it. I mean, it's so flawed. And there are such dangers in embracing these technologies, when people don't really understand the algorithms, the research behind them, I don't even know how you begin negotiating that,
Barry O'Sullivan 23:40
You know, historically, AI was the domain of computer scientists and mathematicians and engineers, and these people, myself included, we wouldn't have been formally trained in ethics, or human rights or legal issues and stuff like that. But this has really in the last five years, become really key. And, you know, you see almost every continent or, in some cases, many countries are developing their own ethics principles, because the public expect it, you know, I think the public notices when does another story in the media, you know, people are watching, The Social Dilemma and the sorts of programs on Netflix and so on and seeing what's going on there. So, I think there is a greater sense of responsibility in developing the technology, there's a greater awareness I think, amongst AI researchers that they have to take these things into consideration. And there's also a greater awareness amongst the person on the street sort of, you know, the company developing the technology that these things are really key and if you if you don't get them right, then people just don't trust the technology at all and they don't use it you know. So actually, yesterday, there was an a meeting where, a very well known Machine Learning professor was telling me that, he bought an Alexa and the first thing he did was, he tried to turn off all of his learning capabilities. Not because he distrusted you know, in some sense but because if he just started with just come to the wrong conclusions about him was quite interestin g. But yeah, there are tensions here that we have to be very careful of I guess.
Gráinne Faller 25:05
Absolutely, actually, I've been meaning to ask you for a while, if you have any advice for people, we deal with a lot of researchers and academics, all of whom are very interested in policy influence, and actually using their expertise to do exactly what you're doing. How did you get to where you are, is, I guess what I'm asking. And we do have advice for anybody who is interested in these things, because we do need people with the kind of expertise than Insight has, you know, in these positions.
Barry O'Sullivan 25:35
Yeah, I think what tends to happen, when you're an academic researcher, you know, like myself, like a computer scientist, is, obviously you spend most of your time in your lab doing your own thing. And if you attend meetings, or conferences they're normally technical, what you find is that the, you know, policymakers who lots of academics are interested in because they make the funding decisions, right. So you can be very self interested and, and think about, well, you know, who are the people who ultimately design the programs I'm going to apply for? But also there comes a point when, you know, we have to take responsibility for the kind of technology we're developing, right, and what kind of impacts it has and what people are saying about it. And so what I started doing was just paying more attention to what politicians were saying about the kinds of things I was interested in, and the kind of problems that they thought were the big challenges. And I found that just talking to them, was an easy door to push on. Because most academics are not politicians, most politicians are not researchers. And they know very little about AI. And so they want to know more about it, they want to have a position, they want to understand what are the issues and what the strong sides of arguments are. And and you just find that, that people are open to these sorts of things. I think you have to be open to accepting to speak and to participate in non technical events. So there are lots of events going on, that are run by legal associations by think tanks, and these are not the typical things that, you know, computer scientists attend, but it's important to attend them. So I think just going out and being a bit more open minded in that respect I think sort of gets you in. So I started doing that about maybe about 10 years ago. And they started, I think another thing that happened about 10 years ago was I realized at the time that's in order to, you know, continue funding my group and my lab at a strong level, I would have to, you know, start drawing down lots of European funding. And to understand how to do that you need to understand how the whole policy thing works. And I found this, I just found that really interesting actually, once I started speaking to people. I suppose the other thing is that you get to meet people who have a totally different perspective on the on the technology. And they ask you questions that you would never have even dreamed of asking. You know, for example, UNICEF, this week have published a report on the impact of AI on children. And it's a really, it's a really great report. And I think people don't tend to think of AI and children, you know, that there need to be specific mechanisms there for children. But of course, there's all sorts of challenges that children sort of raise for us. For example, if you're recommending, content on on Netflix or something, or on YouTube, how do you make sure it's age appropriate? Well, you know, it's actually very difficult, in fact, to make sure the content is age appropriate. Also, things like creating fear in them, by giving them a sense of the thing that they're interacting with, is sort of watching them in some sense, you know, that kind of thing. But also things like, what AI systems should be able to do with children's data and so on, you know, and, and, for example, you take, you know, one of the big application areas, one of the big success stories in in AI is natural language processing and voice recognition. And, of course, children speak and construct sentences very differently to adults do, you have to actually take that into account technologically. And I think, you know, most people don't think about this sort of thing. You do tend to meet people and come up against issues that you might not have thought of, in a particular way. And I suppose also, you end up learning a totally different language, you know, for example, when lawyers talk about AI systems, and there's a big debate in Europe at the moment, because the commission is trying to introduce a regulatory framework for AI, but there isn't even agreement as to what AI is, I think we were saying that the beginning, you know, so we can point to thing over the corner, say that's AI, the thing beside it isn't so easily, you know, it's really interesting.
Gráinne Faller 29:23
It's interesting what you're saying about that diversity of voices, because it looks like even you're looking at the new FIs in Insight 2 in UCC, I mean, there's a big diversity of people from different backgrounds and different expertise areas. I mean, that's pretty exciting. What's your own vision for Insight 2?
Barry O'Sullivan 29:41
Yeah. So in this second phase of Insight, one of the things we've wanted to do in UCC is broaden the type of person who's involved. So when Insight started, we pretty much had straight ahead AI people, people involved in recommender systems, constraint programmers, some achine learners. But now one of our investigators is a Professor of Music. You know, he's, he's interested in the evolution of hip hop. You might say, what the hell has that got to do with Insight and data? Well, you know, he's interested in using data science and analytics to understand well, how hip hop evolves internationally. And so can you identify where European hip hop came from, you know, and that's a really fascinating thing. Of course, it's also just super cool, you know, when you want to talk to somebody, you know, he's, he's a fantastic engaging person. And it's very hard to compete with somebody like that when you talk to the public, you know. And we have people in the health sciences and so on, who are doing all sorts of interesting things that we didn't have before. And we have people in engineering, we've people in mathematics and statistics, which we didn't have before. And also, we've, we've reached out to, you know, old colleagues of ours in the University of Limerick, who've joined, who are experts in industrial applied mathematics and from Tyndall, sensors, all that sort of stuff. So we've completely broadened the group of people that we have involved. And I think that's great, because, you know, you get to cover a lot more application areas or areas of interest. And I don't mean that in a disparaging way, but you know, you cover a lot more domains, and also just come up with questions. Like, you know, if you said to me three years ago that you'll be working with someone who's studying hip hop, I would probably think you were soft in the head, you know, whereas it's fascinating, you know. It's not exactly my area that kind of data science that he's applying to this problem, but you know, you can see there's some fantastic technical questions need to be addressed to answer his questions. That's fascinating. I'd never have thought that that was something that was so interesting.
Gráinne Faller 31:36
Clearly, we've only scratched the surface. So we'll have to have you back, Barry. Thank you so much for talking to us today.
Gráinne Faller 31:47
That's it for another episode of The Insight Podcast. If you like what you heard, please rate, review and subscribe wherever you get your podcasts. Thanks very much to our interviewees. We'll have another great lineup for you next Wednesday. We'll see you then.
Louise Holden 32:07
This has been a Snoring Dog Production on behalf of the Insights SFI Research Center for Data Analytics.
Transcribed by https://otter.ai