The Insight Podcast

Ep1. Colour the past and breathe

November 18, 2020 FH Media Consulting Season 1 Episode 1
The Insight Podcast
Ep1. Colour the past and breathe
Chapters
00:00:40
John Breslin
00:12:30
Emer Doheny
00:20:28
Noel O'Connor
The Insight Podcast
Ep1. Colour the past and breathe
Nov 18, 2020 Season 1 Episode 1
FH Media Consulting

Gráinne and Louise talk Old Ireland in Colour with Prof John Breslin, cutting edge Covid research with Dr Emer Doheny and Insight ambitions with Prof Noel O'Connor.
A Snoring Dog production made by FH Media Consulting for the Insight SFI Research Centre for Data Analytics.

Show Notes Transcript Chapter Markers

Gráinne and Louise talk Old Ireland in Colour with Prof John Breslin, cutting edge Covid research with Dr Emer Doheny and Insight ambitions with Prof Noel O'Connor.
A Snoring Dog production made by FH Media Consulting for the Insight SFI Research Centre for Data Analytics.

Gráinne Faller  0:08  
Hi, and welcome to the brand new Insight Podcast. I'm Gráinne Faller, and I'm Louise Holden and we are going to be talking to all sorts of people in Insight over the next few weeks. You'll find this podcast wherever you get your podcasts normally. We'd appreciate if you would rate, review and subscribe if you like what you hear. This week we are going to talk to John Breslin, and UI Galway.

Louise Holden  0:30  
We're going to talk to Dr. Emer Doheny in UCD

Gráinne Faller  0:33  
and we're going to catch up with Insight's brand new CEO Professor Noel O'Connor with whom this podcast actually all started. But first, John Breslin is an Insight Principal Investigator. He is also a Personal Professor in Electronic Engineering in NUI Galway. He's director of the Tech Innovate and Ag Innovate programs down there. He's also founder of the cultural phenomenon that is Boards.ie. And most recently, he has been nominated for an Irish Book Award for his book, Old Ireland in Colour, which it looks set to be a Christmas hit this year. We asked him how that project began.

John Breslin  1:10  
I suppose initially, I had been doing some genealogical research last year on my family tree and my my mother is from Cla re, my dad's from Donegal. So I was kind of researching both sides of the tree. And as you do that, you start gathering old photographs, I suppose of family members, and I had tried to colourise a particular photo, photograph of my grandmother from Glenties Donegal, when she was younger. And I just, you know, loaded up Photoshop and started tracing out of different black and white parts, and so faces and hair and clothes and so on. I was taking ages basically to do it. So just coincidentally, I you know, my background is Semantic Web in in terms of research and one of the Semantic Web guys I follow, shared a picture of picture that he colourised using a system called DeOldify. And DeOldify is an artificial intelligence  based application for automatically colourising images. And it uses deep learning to do so. So basically, I, I said, I'll give that a go, on the picture I've been trying to do with my grandmother. And sure enough, in a couple of seconds, it has, you know, done a very good colourisation of this picture. So, after that, I suppose I said so you know, and I'll try to send some more images. And I did a few kind of pictures around Galway, and, you know, Irish pictures and so on. And I found that worked really well. And I was sharing them on my own social media accounts. And they were quite popular. And I said, You know what, I need to do a dedicated channel for this, and the Old Ireland in Colour project was born. So that's about August, I think, the end of August 2019. And since then, it's, I suppose grown, I've maybe done probably in the region of about 1000 images, and, you know, the artificial intelligence gets you so far. And then you have to do a bit of human creativity on top of that to to bring a bit further.

Louise Holden  2:54  
Can I ask you, obviously, you saw great value from the very first photograph, did you get an emotional response to seeing a color version of your grandmother?

John Breslin  3:02  
Yeah, because, again, you know, I suppose we all have these pictures of our family when they're younger. And you know, the black and white photographs can be amazing, but when you can see them in color, it just it does bring them to life a lot more. So I think it resonates a lot more and you know, a lot of the historical photographs that I would have colorized they seem to resonate a lot more people when they're, they're in color. But the other part is that the color also highlights things you may not have noticed before. Because there are aspects in the photograph that are somewhat hidden. Maybe it's just because they're in kind of grayscale, or black and white. You your eye isn't drawn to him as much, but then suddenly, when you colorize different features, they become more prominent and more visible. So I think it resonates but also his highlights highlight certain things in the photograph that weren't highlighted before.

Gráinne Faller  3:51  
Can I ask this might sound like a really stupid question. But how does DeOldify know whether something was say green or red?

John Breslin  3:58  
Yeah, it doesn't. So you know, if it's something like grass, obviously it knows it's green, you know, because what it's doing is as what the way it works is basically it's trained on a large image b ank of photographs. So essentially what the creator of DeOldify he's guy in the US called Jason Antic, he would have basically fed it millions of images in color and then the corresponding images in black and white with you know, maybe some noise or stuff added and then it, you know, basically matches the color version to the black and white version says yes, this texture should be this color. So it knows for let's say the common types things like grass and the sea and sky and so on, pretty much exactly what it should be. And of course as you say, you don't know for sure exactly what color something else might be like, you know, clothes or hair color and so on. So it usually makes the best best guess based on the similar types of textures and colors it's seeing and I suppose because people wore all different color clothes. It kind of averages out things. Sometimes like you see in a lot of the automatic colorization stuff that clothes are in blue or a kind of purpley color, you know, some variant of that. So you have to then try and figure out well, you know, is that typical of what clothes color people would have worn? And do I need to go back and manually change that. And for things like uniforms or outfits or, you know, ones where there's historical record, let's say some famous person who wa s abroad, like Oscar Wilde, there's lots of records of the close colors he wore, you will be able to go back and manually correct that. But then for other instances, you're just really making the best the best guess in terms of typical clothes colors at the time.

Gráinne Faller  5:31  
I was really struck by an image of Countess Markievicz, the really famous one of her with the revolver. And funnily enough, in the black and white image, it never even occurred to me that this might be a photo shoot in the studio, which is what it looked like in the color version, and it was kind of extraordinary.

John Breslin   5:47  
Yeah, yeah, you know, that there's a lot of those kind of studio photographs like Markievicz. And, you know, the various people in the US think, because they would have had a lot of backdrops, that's where they would have been, you know, repeatedly, taking photoshoots for various celebrities, and s o on, where the background is a particular kind of thing, like a landscape scene or something. And then of course, when you colourise, the trees, and the the landscape also get colourised , you know, which is probably what it actually looked like at the time, because it wouldn't be in this kind of drab, gray background. So yeah, the one of Markievicz with the revolver, and again, that was one where I would have had to you know, understand the kind of the hue of the Irish Citizen Army uniform she was wearing, which is a particular shade of bottle green, and go back and colourise that.

Louise Holden  6:31  
Tell me about the book itself. It's it's a great big doorstop, of a thing isn't it? How did you go about selecting the images? 

John Breslin  6:38  
Yeah, as I said, I'd done a good few images, you know, it's probably up towards a thousand now. But at the time, when we started off doing the book, it was probably in the hundreds. And, you know, the other thing I suppose about the colorization process is it's improving all the time. So I did all these images last year, and wouldn't have really known very well how to fine tune them or improve them. And also, then in part, the models have improved. So the DeOldify system itself has improved and there's been other AI applications to do things like sharpening and face enhancement, and so on. So for selection, we, you know, we actually only signed with a publisher in the, I'm trying to remember now, I think was the end of April. And so it was pretty rapid turnaround in terms of getting the book on the shelves in early October. It's like, you know, five months. So we basically selected the photographs over that period. And what, we tried to structure them so I worked with an historian in NUI Galway Sarah Anne Buckley, and she was fantastic in terms of helping to, obviously, you know, add the historical context. And she's written the text for the book, but also in terms of groupings, to try and figure out kind of logically how these things fit together. So I would, what I would say is that there's a combination, I suppose between the kind of the, the famous types, and the everyday life. So you know, every like, see kind of the the names of the know, like Parnell, Markievicz, and so-on. But then that's also balanced, I suppose, with the everyday life of people, either in city or in rural situations, which I think is a nice kind of contrast.

Louise Holden  8:05  
So it'll be on coffee tables all over the country, I'd say on Christmas Day will it?

John Breslin  8:10  
I hope so. Yeah. So, yeah, so the book, you know, I've a copy here, obviously, you can't see it, it's 300 pages in in total. And there's 173 images. You know, I suppose something else that kind of struck me was that the print layout is quite different from the screen data. I'm not sure. Is it more forgiving, or what? But it seems to work quite well, in terms of the colors photographs. Like you would look at a lot of them, I think. And, you know, I know, I'm biased, obviously, but I think you would think that they were in colour originally to start off with. And that's a measure of how well the, I suppose the AI process works in terms of colorization. And then obviously, the, you know, that the manual intervention afterwards in terms of making, you know, various things look a bit more realistic. So yeah, there's something about the print versions just seems to work quite well.

Louise Holden  8:54  
Well you're fortunate too in that you have the nice symmetry of all the centenaries happening at the moment and many of the images that you've drawn from are from the War of Independence, period and so forth. So people are attuned to this era of history at the moment anyway, aren't they?

John Breslin  9:09  
Yeah, the timing is good because like obviously, you know, there's there's various documentaries being made. There's one being made at the moment, one has been made about Terence MacSwiney, which is on RTE at the moment, I think it's on right now. And it's, you know, it's around the centenary of his death. We actually have some pictures in the book of his wife Muriel MacSwiney. You know, she went to America after he died and was basically on a tour to promote the Irish cause. So there's a lot of stuff happening at the moment. And you know, I also did recently a series for RTE, which was called War of Independence in Colour. And basically we were taking various images and colorizing those so they would be ones of you know of the the auxiliaries or the Black and Tans or various British soldiers, or, you know, other Irish Independence figures and you know, again, the timing's good. So it's interesting, you know, just to see I suppose the ones, again thinking of that kind of time frame 100 years ago, it seems like a long time ago, and of course, we have seen a lot of black and white pictures, but again, adding th e color I think does add a lot to them and makes people realize, you know how much things have changed. I suppose in some ways as well, because you know, the the idea of these armored cars or tanks running around the streets of cork, or Dublin is quite alien. And so I think the colorization does bring that to life.

Gráinne Faller  10:30  
Will you keep going with it? Or is this is the book sort of going to draw a line under this particular project?

John Breslin  10:36  
Yeah, that's good question. I've had a lot of requests for book two, even though I don't have any plans right now, I think I suppose the best thing to say is we wait to see how this book does, you know, early indications are that it's going to do quite well. And, you know, if it is the kind of one of the Christmas books of the year, that'll be great. But, I suppose it does take a lot of time as well. So you know, I was fortunate I was actually on sabbatical for six months, at the beginning of this year. And the timing, you know, obviously, with lock down and everything, worked out quite well in terms of being able  to do it and have some bandwidth. You know, when you're on sabbatical, you normally have plans to go visit other universities and spend some research time there. And of course, that all went out the window. So I did have a little bit of free time there. And it was a Covid book as well, because the historian Sarah that I worked with, we never actually met and even though we both work in NUI Galwat, we never I don't think we've actually met physically in person. So it's funny to have completed the book with publishers and co-authors and designers and everyone that you've never laid an eye on. And it's great to see that the results, I suppose maybe, you know, maybe the, again just the circumstances, the timing, being offline, we managed to get it done in a much, shorter time. But it's a visual book, and I suppose I had a lot of pictures prepared just in you know, pre publication. So maybe came together a little quicker because of that as well.

Gráinne Faller  11:53  
Extraordinary stuff. The Very Best of luck with it. John, thanks very much for joining us.

Louise Holden  12:03  
John Breslin's book, Old Ireland in Color has been nominated for an Irish Book Award. If you'd like to vote for it, you can find it at Irish Book Awards dot Irish. And we're gonna move on now to Dr. Emer Doheny. Emer, along with many other researchers across Ireland and across the world has really stepped up when it comes to COVID research. And she's been working on an app which uses a mobile phone to monitor breathing in the case of people with COVID. Let's hear more about it.

Gráinne Faller  12:31  
Recently, Minister for Further and Higher Education, Research, Innovation and Science Simon Harris announced 5.5 million euros in funding for 41 new projects through the COVID-19 Rapid Response Research and Innovation Program. Insight's Professor Madeleine Lowery and Dr. Emer Doheny were awarded nearly 145,000 euros of that for their project, Breathe into Your Phone; A new way to monitor COVID-19 patients at home. Dr. Emer Doheny, you're very welcome. And congratulations, first of all,

Emer Doheny  13:01  
Thanks very much. Yeah, we were delighted.

Gráinne Faller  13:03  
It sounds like a fascinating project. Can you tell me a bit about it?

Emer Doheny  13:07  
So the project has sort of two parts. And the first part, the main part is that we want to monitor the way people breathe using the microphone in their smartphone. The second part then is to try to use that data together with other clinical data to predict when a patient may need to be hospitalized or re-hospitalized using using the data that we've derived from their breathing, as well as other information that's currently available through HSE app.

Louise Holden  13:36  
And presumably you're building on on previous research. Is this coming out of another project using the same kind of technology or was this a fresh start when COVID hit?

Emer Doheny  13:44  
So it's a bit of both really. So my background is in sleep and breathing monitoring and movement monitoring using mainly wearable sensors. So that would be like sensors that would stick on to the body or we would attach with tape or maybe like Fitbit watches that we wear on our wrists those sort of wearable sensors. I had already a project going with Vincent's Hospital where we were trying to monitor respiration rate continuously using wearable sensors during sleep. So then when this kicked off in March, April and the funding call came out. We were trying to think of a way that we could apply our research to help. So that was the most relevant work that I had done. But obviously, it would be, we thought it would be easier and better to work with a sensor such as just the microphone in the smartphone to monitor the sound of people breathing, since we wouldn't need to buy any devices, and we could roll it out to a large amount of people quickly. So we're sort of applying the same background knowledge that with the other project, but a new type of sensor with similar type of skills needed from us biomedical engineers where we'd look at things like signal processing, and then how we get clinically informative information out of different types of signals that we're looking at. audio signals, whereas previous projects, we would have been looking at acceleration signals or other sorts of movement signals.

Gráinne Faller  15:07  
So you're looking at recording the sound of people breathing? How will that be helpful, and how is that going to be applied to help COVID-19 patients?

Emer Doheny  15:17  
So the first thing that we want to get is respiration rates, which is how quickly people are breathing, and that would be the most used clinically. So from speaking to respiratory consultants, you know, that would be something they would monitor one on one. They would just watch how people breathe. And that would be the metric, they would write down or enter in their system. And it's a very simple metric really, like just monitoring how many times people breathe in one minute - how many inhales and exhales. It's really easy to monitor one on one, but very difficult to monitor remotely. So that's the first thing we want to do. On top of that, we'd like to be able to look at more detailed ways of monitoring breathing, such as the volume that people are inhaling and exhaling, and maybe variations and how they breathe over time. And that these, these measures may help us to predict a decline in patient's health, but that that's yet to be seen

Louise Holden 16:16  
You're recruiting for volunteers at the moment, now by the time this is broadcast, that process will hopefully be complete. What will be the next step, then?

Emer Doheny  16:24  
Yeah, so hope, hopefully, we will have, we want to look at 100 healthy subjects. We're about halfway there at the moment. And and then the next part is we want to do the same in people with who are COVID-19 positive. So we're going to be trying to recruit patients with COVID, then we'll test our algorithm, we'll develop our algorithms using that data and test it across across those cohorts. Then we're going to call this subset of the healthy subjects and for some lab tests in UCD, as well, where we'll validate our algorithms against research grade equipment. So that'll that'll help us get some more accuracy on the results.  So in parallel to that, we will be working on the predictive modeling work. That's where we will be collaborating with a company called Patient Empower for this side of the work. So Patient Empower are, they're an Irish company that make a patient monitoring app it's a system really rather than an app. The HSE are using it at the moment and they have been since March, when all this kicked off, to allow patients to monitor the way their, their pulse oximetry. So they have a pulse of the finger and the heart rate and patient's temperature, as well as their self reported breathlessness. So we have all those measures over time for a large cohort of patients and we're hoping to analyze that data. And we've a subset of it in the hospitals that we're working with to allow us to predict a decline in patients' health, and add in our respiration algorithms as well.

Gráinne Faller  17:57  
So I assume one of the big challenges at the moment is, you know, if somebody has COVID, and you hear about this, they can become quite distressed, I'm sure there's a big psychological factor too where you're, you know, people are at home, they're sick. And then there's a moment where like, you know, do I need an ambulance? Or should I be calling somebody to get myself to hospital and that kind of thing. That's what this is going to help with essentially Somebody might be able to breathe into their phone with all this data, the technology might be able to give them you know, yes or no, you should probably make the call, or no, you're okay to sit tight for a while.

Emer Doheny  18:31  
Exactly, yeah, it'll either work like that. Or if it was to be integrated into the Patient Empower system, that would automatically happen. So at the moment, their their system automatically has alerts in the hospital, if certain measures go outside of the range that they should be in. If it was to be integrated into that system that would happen. If it was a standalone app, it would be a way for people to monitor how well they are. And people with COVID seem to be very, like as you can imagine, they're like they're worried about themselves and they want to make sure they're well, so people are very, there's very good adherence to using this sort of technology. So yeah, it would hopefully reduce anxiety levels and just help doctors to be able to monitor large amounts of people. They're very overworked as it is. So to be monitoring hundreds of people who are at home and they can't physically see how well they are as well as everyone in the hospital in their normal jobs. It's like it's it's unmanageable at the moment.

Louise Holden  18:32  
I mean, like like many of the of the the research projects that are underway across the world at the moment under awful circumstances, some great leaps forward are being made. I assume this kind of technology will have an application beyond Covid. 

Emer Doheny 19:47  
We hope so yeah, we would we would imagine that we could apply this to chronic diseases that affect respiration. So other like not just respiratory diseases but neuromuscular diseases or yeah and other infectious diseases obviously as well, where remote monitoring becomes even more important.

Gráinne Faller  20:06  
Fantastic. Well, congratulations again on the funding. It's it's one of those projects where you're like, yeah, I'm okay with my taxpayers money going towards it. The very best of luck and Dr. Emer Doheny, thank you very much. And finally, for this episode, it's only fitting that we hear from Insight's new CEO, Professor Noel O'Connor.

Louise Holden  20:28  
Probably many members of the organization assumed you already were the CEO, but you were in fact the interim CEO. And then you are the real CEO? Is that the right terminology?

Noel O'Connor  20:37  
That's right. Yes. It's been a long time, I guess in in preparation, but as of earlier this year, I moved from being the interim CEO. some would say the pretend CEO, to the real CEO,

Gráinne Faller  20:51  
Do you get a crown Noel? A throne?

Noel O'Connor  20:53  
Unfortunately, not no, no, nothing like that. (laughs)

Gráinne Faller  20:57  
We were talking to you back in November. Back when when this podcast was a mere twinkle in FH Media's eye. But we were talking about the plans for insight to and it was all very exciting and lots going on. Of course, we now know that, in events unrelated to the launch of Insight 2, we are currently on lockdown, there's a pandemic happening. Did that kind of take the wind out of your sails?  How have you found the current situation has impacted the plans for Insight 2

Noel O'Connor  21:28  
I mean, obviously the current situation is very difficult for everyone worldwide. And, and it's been very trying, it's been very distressing, it's been very upsetting for many people, and the cost of human life and on human society, obviously, is horrendous. I would say one bright spark in all of that in the last six months for me is how Insight, and more generally, I guess the scientific community in Ireland has responded to the pandemic. The pandemic really gave science a chance to shine, if you like, because ultimately, it's science, which is going to beat this pandemic. And I'm very proud of the role that Insight at all levels has played in the national response to the pandemic. Just to give you a couple of examples, and apologies to anyone I leave out because there are too many exeamples to list them all, but we have some of our maths and stats people working as part of the Epidemiological Modeling Advisory Group, IEMAG, feeding directly into the effort. They've been involved since the very inception of NPHET and continue to work to stay on that, feeding vital information to inform policy on how to deal with the pandemic. We've had people who've worked very closely with the HSE, and the Department of Health on the development of the COVID tracker app, which has been a huge success, possibly one of the best COVID tracker app successes worldwide. And we played a part in that by helping to advise on aspects of the app development, everything from usability of the app, to the ethics and trust underpinning the app and the deployment of the app, obviously, the data analytics aspects and and also the review of the software that was developed, to help cast yet another pair of eyes over over the technology to ensure that is best practice if you like. Other examples include our work on national surveys where we try to understand human impact and attitudes towards the pandemic and things like like lockdown, for example, or levels of lockdown. And of course, our education and public engagement team have been outstanding in a whole variety of initiatives such as contributing to the to the school classes on RTE, for example. And you know, and then of course, there's the stuff that you tend not to hear about, because it's not necessarily in the media or in the public eye, but we've had people from the operations team, going out on weekends with vans full of furniture, to deliver furniture to PhD students and postdocs in their digs or in their shared accommodation, or whatever it might be so so if there's one positive I can take from the change in the world we've experienced it's, it's the, the response we've seen from insight at the very, very human level to help help us better cope with this, this pandemic.

Gráinne Faller  24:22  
It's interesting to see how quickly everything has come together. It must be kind of gratifying for people to be able to respond to a real world situation and to be able to respond quickly and effectively and the way that they have.

Noel O'Connor  24:36  
Absolutely I suppose it's not that surprising, because if you think about it at a very basic level, why do people get involved in research? Because they want to make the world a better place, right. And they typically want to turn their talents and their expertise and their knowledge to important societal problems and challenges. And that's why for example, in Insight we have so much work targeting health, health and wellness, we have so much work target targeting the environment or accessibility in urban environments, or whatever whatever it might be. People want to make the world a better place. So, I suppose the pandemic was a catalyzing factor that allowed people to collectively focus on one big grand challenge of our time. And to do so very agilely and dynamically, quite honestly. But that's because of the energy that they bring to what they want to do.

Louise Holden 25:23  
We've spoken in another podcast at to Suzanne Little and Paul Buitelaar about their research challenge in multimodal data analytics. And they were talking about how the sort of geographical limitations that the structure of insight has imposed on collaboration in the past, but now suddenly, that's everybody's problem. And it's actually been helpful, in a way because now people are working together, and they're more ready to work together across geographies using technology than they would have been a year ago. And so in terms of collaboration and cross pollination, within the Insight Centre, the change has brought you some advantage.

Noel O'Connor  26:00  
Absolutely, I'll give you another very good example that we have our annual scientific plenary coming up shortly. And traditionally, we would have held at one of our sites, and we would have invited Principal Investigators, Funded Investigators to business, the EPE team to one location in Ireland for two days to discuss science. Now, that's obviously not possible at the moment. So we have to hold it as a virtual event. The upside of that is that we save a huge amount of money. And in terms of having to bring everyone to a location, put everybody up for the night, you know, feed and water everybody for 48 hours. And so suddenly, with no budget restrictions on an event like that, we can open it up to the entire centre. So actually, for our upcoming scientific plenary, everyone across the centre, irrespective of role or irrespective of what they're doing, are invited to join our scientific plenary. And that's fantastic. But that's that's a great development, because it really helps build a sense of, of community beyond what we would normally be capable of doing and face to face settings.

Gráinne Faller  27:04  
Do you think this time is going to shape Insight to into the future?

Noel O'Connor  27:09  
And I think it is, I think it's going to change things, I think it's going to change things globally, quite honestly. I mean, if if a vaccine appears tomorrow, right, and it was widely available, I still don't think the world would go back to exactly the way it was pre COVID, I think we've all learned some very important lessons about the need not to always travel to be together in person. And that has huge benefits in terms of people's time, people's availability, and of course, the impact on our environment. So I think we'd see a lot more, a lot more virtual work and a remote collaboration. And I think what this period has proven is that that can be very, very effective, to your point about Suzanne, and and Paul, and how they've been able to work very, very effectively. And so I think that it'll, it'll shape insight in the same way that it is shaping the rest of society. I think the timing is very fortuitous for us. Because I think the other thing that perhaps Suzanne and Paul didn't reflect upon, and it's great, because it's probably already ingrained in their psyche is we have a new research program now at Insight we have a new research program for the next five years that the investigators and the research challenge leads have worked very hard on over the last six months including during lockdown, right. And that is all designed around collaboration across sites across disciplines. So each of our research challenges like RC 4 multimodal data analytics is co-led by two investigators. In that case it Suzanne and Paul, who are from different aspects of that research challenge, you take different perspectives on it, but are also from different sites. Okay. And that ensures that collaboration, and by necessity now remote collaboration is ingrained in all the scientific discussions that we're having. That's one aspect of it. The other really exciting thing that happened. And all of these things have happened, by the way since we last spoke, and most of them since since COVID. broke right, is what we're calling the platform research initiatives. These are clusters of investigators who come together, because they have identified a research topic or a research area that they believe is really really important, and that they want to work collaboratively on. And they basically come, irrespective of site irrespective of research challenge. So it's another layer of integration and collaboration again, they come to work on important topics such as the potential for AI and machine learning in medical image processing -that's that's one example that we actually have, we have six of these already been discussed, and a number of them are up and running already. Another one for example is in the area of cultural analytics, which is about you know, using machine learning natural language processing, linked data to better understand recommender systems to better understand cultural archives, historical archives, so that we can better learn from the past to inform inform the future and that could be things like historical attitudes, of populations around things like pandemics, for example, and what can we learn from previous pandemics that that can help us better deal with this, this this type of scenario and, and similar incidents in the in the future. I would say we have six of those at the moment. And all that's required is for a group of investigators to come together and say, 'Hey, this is an interesting topic we'd like to collaborate on,' and and they get, they get kicked off. So there's just loads of really interesting things happening in Insight at the moment notwithstanding COVID. Notwithstanding that we're all working from our spare rooms or my case the garage that are really exciting, and that I hadn't foreseen when I last spoke to you guys I was talking about Insight 2. So that's been hugely gratifying. I have to say,

Louise Holden  30:46  
That's it for the Insight Podcast for this week. Thank you very much to our guests, John Emer  and Noel. We hope you join us next Wednesday for the next episode, wherever you get your podcasts. This podcast is brought to you by FHMediaconsulting.com. If you would like to comment on the show, if you have an idea for for for a program theme or if you'd like to come on and talk to us, which would be wonderful, you can contact us at info at fh media consulting.com. See you next week. Bye. This has been a snoring dog production on behalf of the Insight SFI Research Center for Data Analytics.

Transcribed by https://otter.ai

John Breslin
Emer Doheny
Noel O'Connor