Instant Genius - Project Discovery: Could computer games help find a cure for COVID-19?

Episode Date: July 13, 2020

In a previous episode of the Science Focus Podcast, we discovered how a team of scientists harnessed the combined power of hundreds of thousands of players of the massively multiplayer online game Eve Online to help in the search for exoplanets. Now, the next phase of this programme, called Project Discovery, is turning its sights from the stars to the coronavirus pandemic. This week we speak to scientists Ryan Brinkman and Jerome Waldispuhl, and Project Discovery’s creator Atilla Szantner about why they intend to turn gamers into citizen scientists to help find a cure for COVID-19. Let us know what you think of the episode with a review or a comment wherever you listen to your podcasts. Subscribe to the Science Focus Podcast on these services: Acast, iTunes, Stitcher, RSS, Overcast Read the full transcription This podcast was supported by brilliant.org, helping people build quantitative skills in maths, science, and computer science with fun and challenging interactive explorations. Listen to more episodes of the Science Focus Podcast: Sonia Contera: How will nanotechnology revolutionise medicine? Chris Lintott: Can members of the public do real science? Dr Erin Macdonald: Is there science in Star Trek? Jim Al-Khalili: Why should we care about science and scientists? Dr Tilly Blyth: How has art influenced science? John Higgs: Are Generation Z our only hope for the future? Hosted on Acast. See acast.com/privacy for more information. Learn more about your ad choices. Visit podcastchoices.com/adchoices

From origin to legend. I need to become a symbol. I need to become the hero. Someone's got to step up. Build the legacy. Let's get to work. You want to get nuts? Let's get nuts.

I'm allergic to nuts. Lego Batman, legacy of the dark night. This is my favorite part. Available May 22nd on PlayStation 5, Xbox Series XS, and PC. Rated E10 for ages 10 and up. Studies. Play. Come together on a Windows 11 PC. And for a limited time, college students get the best of both worlds.

Get the Unreal College deal, everything you need, to study and play with select Windows 11 PCs. Eligible students get a year of Microsoft 365 premium and a year of Xbox GamePass Ultimate with a custom color Xbox wireless controller. Learn more at Windows.com slash student offer. While supplies last, ends June 30th, terms at AKA.m.MS slash college PC. You said this place was steps from the water. We just haven't found the steps yet. How much did we save? Enough.

Enough to get lost. Or you could book a stay with Hilton. Welcome to your ocean front room. Just steps from the water. The Hilton sale is on now. Book on Hilton.com or the Hilton app and save up to 20% to get the stay you expected. When you want savings, not surprises.

It matters where you stay. Hilton, for the stay. day. This podcast is sponsored by name, audio and focal. Streaming has made music more accessible than ever, but true listening is about more than ease. It's about quality. British audio experts name audio, alongside French acoustic specialist focal, combine handcrafted tradition with cutting-edge innovation and high-end materials, delivering digital precision with analogue warmth. So you can experience exceptional sound at home. Music just as the artist intended.

Visit name audio.com to learn more. Anything we can do to help everyday people understand science and help them see that this isn't fake science. This is how discoveries are made. This is how data is generated. This is all that's involved. A better appreciation for what scientists are doing and how complex it is and the work that's done to make discoveries,

I think it's going to lifts up everybody. and especially when you can get them involved. It's just such a fantastic thing. You're listening to the Science Focus podcast from the BBC Science Focus magazine team. With the UK's best-selling science and technology monthly, available in print, and in several digital formats throughout the world.

Find out more at ScienceFocus.com or look out for us in your app store. Hello, I'm Alexander McNamara, online editor at BBC Science Focus. And in a previous episode of the Science Focus podcast, I discovered how a team of scientists harnessed the combined power of hundreds of thousands of players of the massively multiplayer online game, Eve Online, to help in the search for exoplanets. Now the next phase of this program, called Project Discovery, is turning its sights from the stars to the coronavirus pandemic.

This week, I spoke to scientists Ryan Brinkman and Jerome Valdespoor and Project Discovery's creator Attila Shantner about why they intend to turn gamers into sisters and scientists and how they can help find a cure for COVID-19. My name is Attila Thunner, and I'm the CEO and co-founder of massively multiplayer online science, also known as MMOS. And we are a small Swiss startup who came up five years ago with this idea of connecting citizen science with major AAA video games. And so we're kind of the link between the scientific world and the game developers. So I'm Jerome Valdezpul. I'm an associate professor of computer science. at McGill University in Montreal, Canada. And my lab works in bioinformatics, typically we're developing algorithm to study biological data.

And I've been interested for the last decades, actually, in developing also games or programs that allows people to contribute to research and help to enhance the performances of algorithm for the analysis of biological data. I'm Ryan Brickman. I'm a professor in medical genetics at the University of British Columbia. My day job is as a distinguished scientist.

That just means I'm old, I guess, at the British Columbia Cancer, or BC Cancer. And my background is flow cytometry bioinformatics, which is the flow spectrometry data is the focus of the data that we're pushing through today. And so I've got, I'm lucky to have all three of you on today because you are all working on project discovery. I was just wondering if you'd be able to explain to me what, you know, what the project is, where it has been and what the, what you're working on now. I'll start a little bit about the history of project discovery, because this is a story which is going on for five years now. And basically five years ago we started with a very specific idea, which was connecting citizen science, which is crowdsourced scientific activities, with video games, with AAA video games,

was already big, already established player communities. And the motivation was to solve the long-term engagement and activity problems with citizen science. So that was our initial idea. That's why we created our small startup. And we contacted CCP immediately five years ago. CCP was super interested, intrigued by the idea and receptive to it. And we set up project discovery. The first project discovery was together with the Human Protein Atlas.

Then in 2017, we started to work together with the University of Geneva to search for exoplanets in the game. And just recently we started a new project. which involves slow cytometry. And so what was it about that drew you to computer games and made you think, okay, well, this would be a really good way so that we could use gamers to help us solve these scientific problems? So I think it's unquestionable that video games

are the most engaging form of entertainment of the day. We see how people are using more and more video games, how more and more people have access to video games, it's incomparable to any other sort of medium that we digest. Also, it's a very active form of

entertainment. So it gives you a special possibility to integrate something where you want the players to interact with your problem. Basically, what we're doing is we're approaching online communities, in this case gamers, and we try to

smuggle some real-life activity into their everyday digital life. And that's what project discovery is. And you say that there's been two projects so far that you've started on those two projects and you've gone to, what were they, you know, how did they go and how did you come about creating these projects before? The first one was together with the Swedish research team, the Human Protein Atlas. They're curating a big 15 million image database.

It's an open database for researchers of microscopy images of cells and tissue samples. And, well, the first setup was an experimental one. Nobody has ever done this in the video game industry. So we're looking around through our network in research institutions to find good candidate projects for citizen science activities. We try to convince game developers to embrace this feature and implement these kind of mini-games in their virtual worlds. And so the whole process was sort of experimental. We ended up with a very interesting setup where us as a small entity created a sort of middleware

which standardizes scientific data and gives an API to game developers, which has proven to be a really good approach because we could offer this service to other game companies, like recently we started a collaboration with the gearbox. And so after the inception of the idea, less than a year later, we already launched the first project together with CCP. That was Project Discovery Proteins. It was running for one and a half here roughly.

We collected 33 million classifications from players, which is just mind-numbling. It's really unprecedented. Then we switched to another project together with the University. of Geneva, which was introducing light curves in the game. Light curves are these luminosity versus time graphs of SARS, where we ask players to look for dips in this curve, which might signify a transit event, which might lead to a discovery of a new exoplanet around that star.

And there we collected more than 250 million classification. Again, it's completely incredible. And this Monday, that's when we started a new project with this team, we're here today, and whose aim is to use the flow cytometry technology to get a better understanding of it about COVID. So, you know, that's a big goal that we've gone there. The first one is the human genome. The second one, you're just searching for Xoplanes, and now you're using the technology and the gaming.

mechanism that you've got to. What is it exactly? Are you looking for a cure or information or, you know, things that can help produce a vaccine for COVID-19? Maybe I can talk about that because we're on our side, we're on the biology side. So there's a couple things going on. You know, one of the challenges with this data is really complex data. It's 20 dimensions. And the scientists have to go through this data because we limited to looking at data on either on screens, either gaming or the scientists are doing the same kind of thing. And so with these computer screens, we can only see two dimensions at a time of this 20-dimensional data.

And they're trying to navigate through this at two dimensions at a time. And it's a really complex analysis. It's also there's some subjectivity in how the scientists do the analysis because essentially they're drawing circles around dots. And so they can be variability in scientists, and it's also time-consuming. It can take anywhere from five minutes to maybe an hour and 15 minutes to analyze just one sample and to get the answers that they want on that one sample. And if you're doing a clinical trial, you may have thousands of samples just in one trial.

And it is in the news, all these companies right now are trying to develop vaccines and therapies. for COVID. And so there's hundreds of these trials going on at a time. So just masses of masses of amounts of data. And we know from work we've done before with other groups that when they're doing these clinical trials, it can take them up to three months to get the data back from the analysis once they come off the machine that analyzes this. And when people are dying at a very alarming rate, you know, waiting three months for data to come back is It's really not an option. And so by leveraging this crowdsourcing, one of the things we hope is that by putting this

infrastructure in place, both for this pandemic and for the next one. We know this is not going to be the last pandemic we will ever face. That there's a possibility to get answers much quicker, and that's going to be life-saving. So that's one of the early goals. The other thing is, it's a very, again, this complex data set. I said it's 20 dimensions that people are trying to go through. And so scientists are limited in how much of that space they can actually explore, both based on time.

They just want the time to go through 20-dimensional data trying to explore. It's literally hundreds of thousands of different plots on the screen that they have to go through for one sample, if they want to explore their full dimensionality of that. And so by leveraging this crowdsourcing, we have the ability. to really, you know, with 20,000 eyes on one single data set, you have the ability to generate much more data, much more quickly. And with that data, we have the opportunity to find discoveries that would have never been found with traditional analysis because we're just mining the data much more effectively. So we can discover things that are happening in the immune system. We can talk

a bit more about the biology. That would have just been missed. And we can discover it much faster. And the other thing is the data that's going to be generated from this project will help us develop better tools and algorithms to do this the next time around. There's been lots of computer tools that have been developed to do this analysis. But when you're developing these machine learning kind of algorithms, it's really dependent on the training data you have to develop these machine learning algorithms, how well they're going to perform. And there's a dearth, an absence of training data that will help us develop these algorithms. and we're going to generate that kind of data in spades with this project. I mean, there's a lot to take in there. I mean, I guess one of the first things that we should probably establish is, you know,

what is the, you know, for a start, what game is it that people are playing? And what is the game mechanic that they're doing? And then obviously, how does that translate into all of this data, this, you know, scientific data that you can then use? So, well, the problem we try to address here is coming. referred at the clustering problem. The idea is to try to group data into plus that are similar or that should belong to the same entity. So the idea, the game, as it is presented, it's an image when there's plenty of dots inside.

But some region are more dense than others. and the density, the shape or just the shape of distribution of the data tells us if these things should go together or should be grouped into a different cluster, as we name it. And the purpose of the game is we show you this image and you have to separate the region of this image to say this is one group of dots, this is another group of dots. And by the discretion I'm just making here, you understand that it can be sometimes very intuitive. And there's no perfect algorithm for determining if the cluster should be big or rather very small or tight and so on and so forth. So we rely a lot on the visual perception of people at seeing how these things are distributed on your screen. and afterward on the agreement that many players will have together to try to extract the consensus and capture the wisdom of the crowd determining if there is indeed a group of things that should be grouped together here or not.

So the problem as a frame it is very general, and that's why my group has been working on for a while. But I think it was about five years ago, something like this, right? and shoot me an email saying, okay, what you're doing here is cool and clustering, and I think I have the perfect data for you. And we realized at this point, so Ryan explained to me all these work and realized that this problem of clustering is perfectly fitted to the flow cytometry data analysis that he's doing in his lab. So you say flow cytometry. Can you just explain what flow cytometry is? Because that sounds like it's a pretty critical part of this. Yeah, and not a lot of people have heard about it. Even I didn't hear about it when I started my job at the cancer agency.

And I saw what these people are doing. And it's like, hmm, what the heck is that? And it looks kind of interesting. And why are they doing it? Because it looks like something computers can do. So the way the technology works is you sneak up on someone, poke them with a needle and take their blood. And the technology, and the reason why we're using flow setometry to look at these blood samples, it's the word flow cytometry where cytos means cell. And flow is things flowing in a liquid. And so the flow cytometer is ideally suited technology to look at different cells in the immune system and what they are. And so we have, you know, in your white blood cells, you have the cells that you're born with that give you sort of your innate immune system, the immune system you're

born with that just recognizes things that are in your body that aren't you, that shouldn't be there. and helps your body attack that. And the other, they also have the different cells in your blood that recognize, well, I've seen this infection before. I see it again now. It can moment a very rapid response to that. And so there's many different cells that are present in your immune system that mediate or important for attacking these infections, such as COVID.

And a lot of the problems that we're seeing in COVID right now is this huge inflammation in the immune system. They call this cytokine storm. And so your immune system just goes crazy trying to attack this virus. And so with flow cytometry, we can look inside the blood of these people and see what's actually happening at the cellular level, what cells are changing, what cells are disappearing, what cells are being attacked, what cells are activated. And we have to find these different cell populations in this 20-dimensional space because

there's so many different kinds of cells, immune cells in your body. And this flow cytometry technology lets us pull out the specific ones that are changing under different conditions. It's used for cancer. it's used for HIV, any kinds of diseases of the immune system, we're using this technology. Now we're applying this technology to COVID to help solve this problem and to develop therapies. And so that, you know, I guess there's a technology, how that works specifically, you're just collecting the data and then you're feeding that into the game and then letting it loose to the players

within the game to then have a look and see where it looks like there's clusters of data or cell phone. Exactly. We're taking the science. The data is really complex. So as I mentioned, it's 20-dimensional data. And there's a lot of biology that's involved in when the scientists are looking at this data, because the biologists understand that we think in this disease that natural killer cells might be important. And so they're really going to be focused on trying to find this one particular clump of cells

in this huge dimensionality data. And so they've designed very quick ways to navigate through this data to find these one or two or three or 20 populations that they're interested in. Gamers don't understand immune systems and the different cell populations and what's important. And that's okay because this is the cool thing that working with Attila and Jerome has happened is they don't have to because we have the power of the crowd. We don't have to shoot them through and explain to them what path they have to navigate through this 20-dimensional data because we have so many people.

We just say, here's everything. Find everything in all possible combinations. And then we'll sort it out later. And so we'll give them all the different projection of this data to look at. And they'll find everything. And so they'll find the stuff that scientists want to find. They'll find stuff that is probably not important because it has nothing to do with disease. But the cool thing is they'll find stuff that scientists might not have looked for because of time or because they're looking where the light is.

And we'll have the chance to discover this. Maybe if I just want to top on that, I think that's very important what Ryan just said. because through crowdsourcing, we'd be able to have a look at this data as it has never been done before. And potentially what we hope is like we find a better way to navigate these data to be able to extract its full potential and eventually realize the promises that all this technology can deliver to us. So I guess it's a case of that when you, when you, you collect the samples and the data, all the data is there, you just have no way of filtering through it and you're using the power of the people who are playing the game to be able to

filter that for you in a way that you hope, I guess you hope an algorithm or an AI would be able to do in the future. Exactly. And there are lots of algorithms that have been developed so far today. And they're really great for doing discovery kinds of problems where you say, I want to know what's different between all the sick people and the healthy people. But when you try to apply those same algorithms to do clinical reports to say, I'm really interested in finding NK cells. Because these algorithms are for the most part unsupervised, which means they're not giving a lot of information about the data that they're looking at when they're trying to find these clusters. And so they have to adjust for the size, shape, and distribution of all these different

cell populations in very high dimensional space. And all these cell populations, look a little different in how they're stretched out or squished. And so these unsupervised algorithms really, they don't have the performance when you're trying to do clinical reports. And so they're not really used for that. But human eyes are, our eyes are designed to find patterns. And we've, you know, we've evolved to detect patterns so we don't get eaten by tigers. We're trying to, we're trying to apply this really awesome pattern recognition that humans have to data sets to help build better algorithms to learn how humans have been trained over millions of years of revolution to find these patterns and then say, hey, look at what these humans are doing. This is how they find

these patterns and show that to computers and go, oh, I get it. And then they'll go off and do the same kind of thing. Presumably for that to work, you need a significant amount of human eyes looking at that data. Yes. We need lots of data and we also need them to do a good job. Because if the humans are gaming with one hand and shooting space aliens, while at the same time they're trying to draw clusters around dots, we may not get the best data. So we need lots of data and we need lots of good data. So one of the things that CCP and the groups there have been involved in teaching the lay people and the gamers,

how to do the job that scientists are doing. And that's really interesting. So how do you go about choosing a game to feed this mechanic in? And then once that's in, how do you go about making it accessible to, as you say, people who are shooting starships to then go off and be scientists, as it were? So, you know, in the last five years, my big part of my job was to go to gaming conferences and events and talk to game developers and publishers and convince them that this is something super important. This is amazing to do. This is a lot of fun working together with scientists, bringing. in this to the player community,

and it brings a lot of value to the game. And so mostly we're aiming for bigger games, because as Ryan was saying, we need a lot of players, a lot of activity in this citizen science project. In general, citizen science is based on this idea that we have to hand out the same task to many people, and then we do some sort of aggregation

or major devoting to see who are the guys who are often, and two other guys who are doing a good job. And this way we can provide very high quality output. So we are looking for bigger games with potentially hundreds of thousands or millions of players. In the case of Eve online, Eve has roughly half a million active players every month. So it's a big game.

We have a big player community who can contribute to this project. Not to mention that Eve is... You know, Eve players are a crazy, crazy, interesting bunch of people they say that 99% of them are science buffs and they're used to solving very difficult problems that love to be challenged and they really welcome

the project discovery. The other project that we launched recently was together with Borderlands, which is a completely different kind of game, a bigger game but a much diverse audience and again there we worked together with Jerome to bring another scientific problem there. But that's one thing. Also we have to find places in the gameplay where such an activity works. So, for example, in Evo Online, there is a lot of idle time.

You're waiting for your buddies to come with their spaceship to fight. And while you're waiting, you just open project discovery and solve a couple of tasks. And these tasks take, I don't know, 10, 10 seconds, 20 seconds, something like this. So every time you have a little bit of free time, then you can contribute to science and a process. This might not work that well in another game like, if you should. take League of Legends while you're playing you know you need 150% of your attention there's no way you're going to solve scientific problems while you're playing although you know if you take these um mobile games like League of Legends there is always this network

synchronization part where we're waiting for half a minute two minutes for other players to load the client it would be a perfect place to add some scientific activity to to contribute something really meaningful in the meantime So yeah, yeah, yeah, these are the games that we're looking for who can provide massive crowds. And I guess the hope is that the same enthusiasm that the even line crowd players took to the previous discovery missions, they will take that to the new one, especially, I guess, given the fact that there's a global pandemic on at the moment. Yeah, I think this is an extra layer of motivation, so to say. simply because the situation of what we are in as humanity.

In general, it's really interesting because there has been extensive surveys of why people participate in the first place in any citizen science activities, not just in games. And these surveys shows that the primary motivation is to help scientific progress. I think it's super important because that's a very solid foundation for this project, that people want to have. There's an intrinsic motivation. Of course, in games, you know,

putting or taking away this effect, what we have with COVID right now, we have this additional layer which is reward, in-game rewards. So we're very attached to your favorite virtual world that you're playing with. And we are sort of thanking players for contributing with in-game rewards,

which connects this whole activity to the bigger picture to the lower the narrative. So there's a dual psychological effect of that, not only helping the planet, you're also getting some in-game rewards out of it as well. Exactly, exactly. You know, it was really interesting because with the first project discovery, CCP game designers created a new in-game currency for just project discovery, which was called Annelesis Credits. Now, you could use this analysis credits to go to a special shop, buy something and then sell it on the market, and get the main In-Game. game currency, which is risk. What we've seen, there were many very active players who never bother to spend this analysis credits, which is a clear sign that they were not doing it for the

reward, but rather for, because they wanted to help the science. And the uptake has been incredible. When they, when they, when we start talking about how much, how many answers we could get, I was like, really? And I wasn't sure. And I think so far we've exceeded expectations. I think the number you guys have a bit more answers and the numbers. I think we had like 500,000 puzzle solves on the first day. I think it's even higher.

It's crazy. It's insane. So as a scientist, you must be I was just going to say, as a scientist, you must be really, really thankful with this huge amounts of data that's coming in, like almost instantaneously. Actually, we're a bit worried that

we're running out of data. And so my team is basically trying to get more data as we speak because because we have to keep that bucket full. They're just going it through so fast. We thought, oh, that's going to be okay. We have a couple days to generate the next round. It's like, oh, my gosh, we're running out. So I guess this is great news, obviously, for you to have all this data.

So what can you do with this now? You've got this data now. What can you do with it now? And how long will it be before you can actually say, okay, this is the science coming off the back of, you know, players of a game doing some spare science in their time? Science is a long road, and it's frustrating for people. It's like, why don't we have a vaccine yet? Just to design these flowstometry studies takes weeks easily,

and then you have to go out to the patients and poke them with needles and collect all their blood, and you have to get ethics approvals for what you're doing. You have to organize scientists and clinicians. So just to generate this data is a massive amount of time. And so a lot of these studies are just starting up. So we have this infrastructure in place and some of the data is starting to filter in, but these are the first studies.

We now, we have data from the various first outbreaks in Italy. That's the data that we're getting today. And that story was weeks ago, a couple months ago, really, right? And that data is just now filtering into our system. And so a lot of these clinical trials, we don't have the data yet. They're just getting up to speed. And once we put this data through the system, as Attila has mentioned, we have to massage that. We have to take hundreds of, 100 players analysis of the same data set and figure out what the best way to put that together to get the one best answer.

We've never done that before. So we're developing the algorithms at the same time as the players are generating the data to help us understand what they've done. And the next thing was we have to start looking for, and after that's all done, and we have to tease apart these massive data sets to try and find out what's changed. and also trying to overlap that with what the stuff the scientists has looked for. So we're not going to get any answers next week and probably not even next month. Very early on, we're going to be able to give the scientists data to help put that in their hands as soon as quick as possible. So they can put their expertise on this soon. So one of the big ideas we have from this project, it's not just the people that we have here involved.

They're looking at this data. All this data is can be made available for everybody to look at in the community. other scientists, not just ourselves. We're going to expose everything in an open science kind of way. So we're crowdsourcing the idea is we're going to crowdsource scientists to look at this data that we've crowdsourced gamers to look at. And so hopefully with more eyes, more scientist eyes on this data that more gamers' eyes have looked at, we can get to the answers quicker.

But it's science. And we don't know when that next discovery is going to come. Yeah, can I just add on top of this? It's like what is very interesting, I think, in the project as well, as you can guess from what everyone mentioned. Like, there's different levels. I mean, in the sense that there's the urgencies of finding something for COVID and things like this. And we're trying to build the technologies to advent this. But we have to get prepared also for the next outbreaks, potentially.

And all the technology and the infrastructure that we're building here potentially can be reused afterward to apply this technology much more faster for using flow satrometry at a large scale on emergency users. And so here we're speaking about COVID, but there can be another source later. and Ryan's using that for cancer patients. So it can make a breakthrough in terms of how we use this technology in day-to-day. But this is a general problem with the scientists. You know, everyone's going to ask, when is the cure going to happen and when's the next discovery? And I think no scientists would ever tell you, oh, we're going to have the answer for you next week. It's a science.

If we knew how it was going to be done, it wouldn't be science. Science is about discovery, right? And so things can happen very unexpectedly. You know, you look at that bread on the table and see there's mold growing and then boom, you're done. And sometimes these things take years. We're looking for that mold in COVID. And there's some things that we know to look for and we hopefully can get answers very quick. But it's not going to be by the time this podcast goes live.

And this is why basically what CCP does this is really groundbreaking. and changes the way we can potentially do science a large scale, right? As Attila was mentioning earlier, it's not the first project they're engaging, but they really build different infrastructure. How can we accelerate science by engaging all the communities of gamers there? That turns out to be very enthusiastic for science and very skilled in many complex tasks. Yeah, it seems like this is something that, obviously, that the work you're doing is groundbreaking,

but it seems like something that could be applied to lots of other things, like going forward. And at the moment, you're using this to, you know, you say you search for exoplanets and now you're searching for COVID. But, you know, what other applications could we be using this crowdsourced citizen science for what other things are out there to be,

you know, that we could put this application of citizen science towards. Besides other applications, full spectrum data is used everywhere. It's for cancer. for diseases of the immune system. We have cures for cancer now that have to be developed one patient at a time. And so aside from COVID, there's just so much false photometry data that's affecting so many people. Anybody use leukemia or lymphoma. It's everything that's shot through

photosochometry. So the technology has been put in place for COVID, but there's so many diseases are affecting millions of people that we can apply this technology for, aside from the citizen science stuff that Jerome and Attila can talk about. Yeah, I think generally speaking in life sciences, we see a lot of cases where data is relatively easy to acquire, or easy is probably not the good word because this is really a bleeding edge technology that they use. But let's say we can acquire very large data sets. On the other end of the story, we start to have very good machine learning algorithms. But as Ryan was explaining, we need that link to that.

So we need training data to make these machine learning algorithms really effective. And that's what players can provide. So we can feed these big data sets that comes from life sciences and then improve these machine learning algorithms for life sciences. Yeah, just to basically emphasize indeed what Atila said. Now we see we have the emergence in last years about the broad use of machine learning and AI in many different fields. these are very powerful technique, but they always rely on a large data set that can be trained on and potentially, I mean, as large as possible to make the technique reliable. And this is a very timely project that we have to put on, because for many problems and complex problems in life science, indeed, we're missing this data that needs the workforce of,

of a scientist spending hours doing these things by hand at a very small volume. And the infrastructure that was deployed by project discovery basically enable us to scale this, to scale this up to the level of very large communities and really unlock the potential of artificial intelligence techniques. So basically any time a scientist is looking at a picture and is doing an interpretation of that,

that data is applicable to this crowdsourcing. So X-ray images or anything like that, the same kind of ideas apply where the human is interpreting an image to get an answer. Those are complex problems for algorithms to do. And this is the kind of thing that humans can help train those algorithms. Do you think we'd ever get to the point where the algorithms would start to be able to learn from that and get better themselves without having the need for quite as huge number of players to be able to look at the data? I think trying to identify in Instagram pictures, if the person is holding a cat or a dog,

the kind of information that you use to make those kind of decisions is very different than if you're looking at a flowstotometry data, we're trying to do dots in space. And so you really need to have algorithms trained for the type of data that you're looking at. Otherwise, it's not going to perform as well. So, you know, obviously, you know, science is a long process and it takes its time. The outbreak of coronavirus has been a recent thing. The point that I was looking at is obviously you've got this technology now. Will you be able to swiftly move that on, you know, if the data changes?

So if, you know, we discover something that we haven't known before about, you know, the COVID-19. Will you be able to adjust what you're doing to be able to sort of hone in on actually this is what we need to target on? I think this third edition of Project Discovery was very interesting from that perspective. From the very beginning, CCP had this approach that they wanted to create sort of framework inside the game that is capable of taking citizen science micro-task and showing it to players. because they knew that, you know, it doesn't matter which project we start with, will eventually change it to something else. So in the game, we already have sort of an infrastructure.

As I was saying on our side, we, from the beginning, created an API for grain developers, so that piece of software is continuously improving, but that's also in place. So right now what we've seen, and that was quite amazing, that basically in a couple of months' time,

we managed to bring it into the live game, from having the idea of using flow cytometry data to help these research efforts. And I think that is something very, very special. And not only did they develop the interface to it, they made it look really cool. And so that's important. You know, if you just have a boring, you know, run-of-the-mill interface that scientists are used to, it's not going to be engaging. But they made it look really cool.

And so if you're in the game, it's like, this is fun. They make it fun and interesting, and they tie it to the game. Just the way they present the pallets of colors and the extra graphics they show on the side. They spacified what scientists doing in their cold laboratories and made it look like a game. But it's doing the same kind of thing that the scientists are doing in the laboratories. It's just amazing to see. Is that kind of the key to making these citizen science projects work, making it seem, like really natural for the environment that they're in.

Yes, exactly, because I think that's the power of this whole setup. I mean, this is why we started in the first place. We knew that if we managed to do something like this, this sort of very organic integration. So this task becomes part of the game, part of the virtual world that you love so much, that you spend so much time in, that is connected to the lore and the reward system,

Then we can have that kind of activity, what we see actually in project discovery and these citizen science project. They even took one of the scientists that developed the data, Andrea, our collaborator in Italy, who provided some of the COVID data. They actually took him and made him an avatar in the game, and he walks through the players teaching them how to do the analysis. And I saw that. It's like, peace out, I'm done. That's the coolest thing I could ever be involved with. It was just such an amazing thing to see. And so they're really integrating the science directly into the game.

You're having the guy who's a frontline clinician who's working on COVID data teaching the players how to do his job. How cool is that? Yeah, it's really interesting what Ryan was saying because, and I think that's an important part of this project is that it's not just about the data because the data is all really super interesting and very valuable for research, but it's a unique opportunity to do science outreach. Just think about this project.

I mean, Ryan said that even he didn't know about flow cytometry when he started to work at the lab. Now we have potentially millions of people in the game and through these interviews who learn a lot about flow cytometry. Why is it important? And this kind of knowledge or this kind of science outreach helps these research efforts because then they have much more support from the general public. And I see with all these project discoveries, we always had researchers very involved in the process.

For example, the last one with Exoplanet research, we were honored to have Michel Meyer to be the face of the project. So he also got an MPC in game, a non-player character in game. But he also came in person to Fennfest, which is the annual player convention in Reykhevik, where we have 3,000 or 4,000 players coming every year. And he gave a presentation about how he discovered the first exoplanet, you know, for which, by the world, he got the Nobel Prize last year. And I think this thing that a researcher of his caliber is coming to a gamer convention talking about his research, I think it clearly shows that scientists taking this opportunity

very, very seriously. And anything we can do to help everyday people understand science and help them see that this isn't fake science. This is how discoveries are made. This is how data is generated. This is all that's involved. A better appreciation for what scientists are doing and how complex it is and the work that's done to make discoveries,

I think it's going to lifts up everybody, and especially when you can get them involved. It's just such a fantastic thing. That was Attila Shantner, Ryan Brinkman, and Jerome Valdes-Ball, talking to me about project discovery. Let us know what you thought about the episode with a rating or review wherever you listen to your podcasts, and check out our previous episodes at ScienceFocus.com

forward slash science focus podcast. The new issue of BBC Science Focus magazine is out this week, and in our cover feature, we look at the missions planning on building a permanent base on the moon by 2030. As ever, there's loads more inside, so head over to the website to find out how to subscribe. Goodbye. Thank you for listening to the Science Focus podcast

from the BBC Science Focus magazine team. We're the UK's best-selling sites and technology monthly, available in print and in several digital formats throughout the world. Find out more at Science Focus. or look out for us in your app store. This podcast is sponsored by Name, Audio and Focal. The texture and emotional depth of music can be lost through digital sources or poor signal.

Name Audio believes you can have digital precision with analog warmth. Alongside French acoustic specialist Focal, Name creates high-end audio systems combining innovation with craftsmanship so you can listen to music, just as the artist intended. Discover more.

at name audio.com. Ambition comes in all shapes and sizes. At First Citizens Bank, we roll with your goals because we're built for what you're building. Fit for your ambition for Citizens Bank. Ryan Reynolds here from Mint Mobile, the message for everyone paying big wireless way too much. Please, for the love of everything good in this world, stop.

With Mint, you can get premium wireless for just $15 a month. Of course, if you enjoy overpaying, no judgments, but that's weird. Okay, one judgment. Anyway, give it a try at mintmobile.com slash switch. Up front payment of $45 for three-month plan, equivalent to $15 per month required. Intro rate first three months only, then full price plan options available. Taxes and fees extra. See full terms at mintmobile.com.