Instant Genius - Project Discovery and its search for exoplanets

Episode Date: March 29, 2018

We talk to Bergur Finnbogason, Development Manager for Project Discovery, which uses players of the Massively Multiplayer Online game EVE Online to help search for exoplanets. 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. You said this place was steps from the one. 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. Study And 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 Slaternity offer While supplies last ends June 30th

Terms at AKA.m.m.S. College PC 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. The problem is that, you know, it's space and there's a lot of stuff flying around. So oftentimes you get like, you know, incredible noise. And a lot of the noise, you know, it's not, doesn't really have this frequency that we're looking for. And this really messes with algorithms.

So this is where the human eye really helps. You're listening to the Science Focus podcast from the BBC 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 and welcome to the Science Focus podcast. I'm Jason Goodyear, commissioning editor of BBC Focus magazine. Currently, astronomers are looking for exoplanets circling neighbouring suns.

But instead of pointing our telescopes to the sky, we're entering the world of Eve Online, a massively multiplayer online space fairing game. Science Focus editor Alexander MacDomara, speaks to Berger Finn Borgesen. Berger is the development manager for Project Discovery, which taps into this fast gaming community to gather data. He explains what makes gamers so good at searching for exoplanets

and how soon it will be before we discover a new planet using computer games. Would you be able to tell me who you are and what your role is and what it is that you're doing with Project Discovery? Yeah, absolutely. So my name is Berger. People call me Berger, because that's easier, I guess. And I'm a development manager here at CCP games in Reykjavik. I mainly work on an MMO called Eve Online, MMO being massively multiplayer online game.

And Eve is, yeah, I mean, Eve is a space MMO came out in 2003, so returning 15 this year. And a couple of years ago, or three years ago, we got in contact with a Swiss company called MMOS. and they had this idea, this hypothesis, that they could actually like, you know, take citizen science and, you know, if they implemented it with MMOs and gaming, they could actually, you know, start to see real significant impact in science and citizen science. So two years ago, we actually implemented the first version of product discovery where we worked with the human protein atlas in Stockholm, Sweden. the idea there was to help HPA or human protein atlas to categorize their subcellular subcellular protein atlas so we basically had you know hundreds of thousands of microscopic images and that our players would categorize down to like you know the proper categories there were around 23 four categories and the idea

around this is that these problems are problems that are just a wee bit too difficult for computers to solve and they need kind of human insight to really, you know, crack the code. So that was the project you've done before. So what is the next one now that you're doing? Yeah, so we, so the first version of project discovery ran for a year. And we got that got really good data and incredible, incredible, incredible. results from it. And so a year ago, we actually launched the second iteration of project discovery called Project Discovery Exoplanets, or we codename in that. And we're working with the University of Geneva on that project.

And that project is all about searching for exoplanets or planets orbiting all the stars that we see in the sky. And that, I mean, that's definitely like the ultimate science project we wanted to get into the game. It like fits really nicely into the theme of the game and kind of the lore of the game. But we wanted to make sure that like, you know, we wanted to take a different project first to really kind of harden our hypothesis, our game design, our tech. So once we got this, you know, the dream project in, you know, it would all be like super smooth

and super smoother and good. So how is it that these, the people who are playing, even on, how is it that they're able to help you discover exoplanets that are out there in the solar system. Not the solar system, obviously, the YD galaxy. Yeah, so actually what we did is, project discovery is basically a mini-game within E1 line, and you can launch it wherever in space. And for the exoplanet project, actually, you fire it up, this system,

and it has a really strong in-game connection, so it's tightened with the law and kind of, you know, the progression of the game. And what you're presented with is these light curves. So they kind of look like, you know, very spiky graphs. And what you're looking for are transits. So these light curves are a series of observations of distant stars. And every time something passes that star, you get a – you basically get – get a transit, you get a dip in the light curve.

So what the players are looking for are these dips in the light curve. And if they find them on regular intervals, you can, you know, with some certain, certainly say like, hey, you know, is there a plant there? There is something there blocking the light. So that's basically what the players are doing. But the problem is that, you know, it's space and there's a lot of stuff flying around. So oftentimes you get like, you know, incredible noise. And a lot of the noise, you know, it's not, doesn't really have this frequency that

that we're looking for. And this really messes with, you know, algorithms. So this is where the human eye, you know, really helps. So why is it that, so when you say the human eye is better, what is it that makes the human eye better at recognizing these patterns than artificial intelligence or machine learning? Well, humans are really good at pattern matching. And, you know, we are, we really long for patterns in all things.

I think it's just, I mean, it's something that's just so ingrained in our being, I guess. Yeah, I think so. That's my hypothesis. Now you're going to fact-check me and it's all like wrong and, oh, no. I'm sure it's fine. But do you know, is there something about the Eve community that makes them particularly good at spotting these differences in patterns? yeah I definitely think so I mean the eve community is a

very dedicated community and I mean I would say that they are I mean it's a it's a bunch of super smart players and incredibly intelligent players so I mean that's definitely one of the learnings from the first version of product discovery

I mean we were quite scared in the first product discovery that it would be too complex or too convoluted too awkward and we set ourselves you know some rules when we when we pick projects or when we pick projects for product discovery and one of the things that that we added to these rules

actually after the first one is that the project that we pick needs to fulfill a certain level of complexity we don't want to you know undermine players capabilities or like you know simplify it and and kind of dumb it down so we need to maintain like a certain certain level of complexity so yeah in theory yeah maybe they are they are probably smarter than the average human being. They're definitely smarter than me. Eve itself is set in space.

So do you think that that gives the players already the sense that they want to be looking for things in space and other planets and all that? Oh yeah, absolutely. I mean, I think this, I mean, we make a space game and most of us are complete space nutters and working with exoplanets just, you know, emphasises that.

So it definitely like tickles our nerd. And when we're excited, that's a good indication that our player base would be excited. Do you think that when a player sees one of these sort of dips that they can say, hey, I've found an exoplanet. Can you then go and say we've found an exoplanet or what happens with that data after that? So actually, exoplanets are super, super, super rare. So what we do is it's really hard, you know, when you're crowdsourcing kind of science data, you need ways to measure people's accuracy and people's efficiency in doing it. So every, you know, third to every tenth image or curve that the player gets is actually, you know,

it's a solved curve, it's a controlled data that you get. And we use this controlled data to messier people's. accuracy and success. So in the control data or the gold standard data, we have a relatively high amount of light curves showing up to basically

help players recognize these light curves. And also just, you know, it's more exciting to find something rather than, you know, just find nothing. And this is one of the big design challenges for the second iteration of discovery is like, how do you make a fun game

about finding nothing? Like, In a way, like, what's the fun in saying like nothing here, no transit? So it was an interesting exercise. But on, so once we find something, we've actually, you know, we have found some promising light curves. We basically, these light curves get packaged up and sent to University of Geneva for further analysis. And that's basically what's happening right now is that they have a good bulk of light curves

that they, you know, they need to look further into before. they can confirm or deny that there's an exoplanet there. And we were actually getting, like, the feedback we're getting from University of Geneva is that we were actually finding stuff that they hadn't noticed before. And, you know, players are marking in, you know, strange things that they hadn't noticed or they hadn't paid attention to. So there is actually a lot of good feedback.

They're getting a lot of good feedback from our community. So how long will it be before we can say there's been a new exoplanet discovered out there and that was discovered thanks to the EVE Online community. I hope it's going to be very soon. I'm super excited. I've been waiting since the game came out for that news. And do you expect it to be soon or do we have to wait a few more years for this sort of thing?

Well, I hope it's going to be soon. Like up to today, like from when we started until today, we've actually had 65.7 million submissions. So 65.7 million light curves have been examined. For the light curve to actually reach consensus, you need between five and seven similar submissions or submissions in certain range on the same light curve. I mean, we have close to half a million players

that have participated in this version of prototype discovery. So, yeah, I mean, we've definitely submitted a lot of data, so I hope we'll get to see something soon. But then again, it's very, very rare to find exoplanets, and it's super exciting. Do you think that the exoplanet that will eventually discover might get sort of thrown back into the Eve Online universe or even the other way that Eve players get things named after them?

out there. Yeah, so we're actually working with, so one of the rules that we set ourselves when we're picking data is that it needs to be like thematically fitting for the game. So we think immersion is a critical factor when it comes to like getting people excited and engaged with this project. So what we did in the previous project and this project again is that we got one of the one of the professors, one of the leads of the projects in real life and we emily

them in-game and we actually created the in-game character. So we took an astrophysicist called Michel Mejor, who is kind of heading this project and we actually made the in-game version of him. So he's this kind of big professor for one of the big corporations in Eve

and he's basically kind of championing the project within the game and kind of he's kind of the backbone of the in the game. We got Mr. Mayer to come to speak at Fanfest last year, and this was one of the questions

our players brought up. And he had a very good scientific reasoning why we should not name planets, unique names, because they have this super good efficient system to name them, so everyone just knows right away

what planet you're talking about. But I mean, the other way around, we should definitely look at that. Find some cool planets. We should definitely look at creating some interesting connections in our universe. And for these players, so obviously, you know, it may go back in, but what about what do the players get out of themselves other than the fact that they might have helped spot

an exoplanet? Hmm. I mean, of course, like in the game, they get some in-game rewards. And what we found with the first project discovery, I mean, we could kind of categorize the players into three groups. So the reason why players played into three groups, the first one being just, you know, they're just doing it for the reward. Second one being, you know, we're doing it for science. And that was actually a really, you know, good chunk of players that are playing this for science and, you know, contributing to science.

And then the third one is just, you know, I want to be the best at this. I want to be the number one player. I want to be the greatest. It doesn't matter how many hours they need to put in. Like, I want to, I want to be the greatest art prod discovery. want to nerd out on the forums telling people that I'm like beyond them in spotting exoplanets and you know. For the, you know, the hardcore game is that one.

The hardcore gamers, yeah. I can understand with the people who are really into the science, just like those are the ones that are going to be really pouring over it. But how do you sort of cut out that noisy data, which might be people just going through for in-game rewards or just wanting to get to the top of the pile? Yeah. So we do that with, we calculate players accuracy. and the higher the accuracy,

the fewer gold standard image, like data you get, or kind of the less controlled data you get. So we actually, we actually, we use the accuracy a lot there. So once you hit below around 50%, you stop getting actual data. You only get gold standard data.

And once you go below, I think it's very, 40% so you stop receiving awards altogether, rewards altogether. So we basically just kind of slowly, you know, either push you out or, you know, reinforce that you should like, you know, get your act together and, you know,

up your accuracy. And you start, when you start the game, you basically run like a tutorial. And in that tutorial, we can fairly quickly determine how efficient you're in the game. And, you know, if you are, you know, if you're actually going to graduate to real data, or if you're just going to kind of float out and leave the game.

So that's one of the things. That's basically one of the levers that we've had to play with through the projects to really figure out what's the best moment to actually let people into real data. And did you find that as the more people worked on this data, they got better and were more accurate? Yeah, I mean, yes, definitely. We have, I mean, our community is super engaged. And we have, I mean, they're very active on, on, you know, Reddit, you know, media like Reddit and so on and so forth.

So there is a lot of knowledge here going on. And we saw in the first product discovery, we actually had one of the in-game universities called EVE University, which is completely run by players. They actually hosted a couple of seminars on protein analysis. and, you know, it was really popular with players and, you know, players were actually reading up on science papers from HPA. And so this time around, what we did is we actually worked with the University of Geneva to make a MOOC or like an online course on Exoplanets and Axoplanet and Exoplanet.

And they've seen a healthy Arctic in those courses and, you know, kind of alongside Prodiscovery coming online. So that's a really interesting and exciting thing. So in theory, like as players are kind of going deeper, they are also, you know, looking online, they're Googling stuff, they're going to these courses, reading papers, so they're becoming better and better. And, you know, some of the arguments and comments and feedback that I'm getting from Project Discovery now, like, I mean,

some of it I don't understand because they're just way too deep in the science of things. So there's clearly an engaged audience for science. this sort of citizen science project, are we going to see more of this in the future, especially built into games like Eve or other games that are out there? I definitely hope so. I mean, it's as a joke, after we wrapped up the human protein Atlas work,

we actually calculated around 146 man years worth of work that our players contributed with their, analysis. So you can definitely see, like, yeah, okay, scientists might do it quicker, but yadri, yad, yari, but the thing is like, you know, if we are, if we can actually free up scientists' time to, you know, focus more on the results and, you know, analyzing the data versus just categorizing and, and doing, you know, this kind of more kind of ground level stuff.

I mean, I think that's the way going forward. and I hope that that will actually like allow people to to unleash their powers and their science powers. So how do you like when you're coming up to a project, how do you decide which project it's going to be? So for example, you say you've done the proteins, you've done exoplanets now. Where's next? Yeah. So we basically set ourselves some rules when we started working on this. And the first one being that whatever project we pick for project discovery or whatever data we pick for product discovery,

it needs to be of altruistic nature. So, you know, yes to figuring out how to, you know, cure cancer, no to looking for, you know, viable targets of war or whatever. The second rule being it needs to be thematically fitting for the game. We think immersion is a critical factor in, you know, in getting people emerged. And this is one of the hypotheses that MMOS had coming into approaching us. they were looking at all the citizen science projects out there

and one of the biggest issues they were facing was extremely low retention so you had a lot of people come in but they would disappear as quickly as they would come in and MMOs, no MMOs massively multiplayer online games I mean we have kind of to a point we've actually solved

this issue or we are further down the line of solving this issue so merging the two can we actually can citizen science benefit from the successes of MMOs. And we're actually proving that that's definitely the case. But we look at the immersion as a critical factor in there. So the data sets also need to be of critical scale or critical size.

Our players are super eager to take on challenges, and they actually run through huge datasets really quickly. with the protein data set there was around 250,000 images and we managed to reach consensus or full consensus on the whole data set within three weeks and we had kind of

we were thinking it would take around three months but it just really took off and left us in a tough spot the fourth rule is the data needs a consistent time to solve so we want to make a clear promise to our players

when they open the game, we need to be able to promise them that this will be a 20 second, a one minute or a five-minute thing. It can't be like one moment it's 10 seconds and the next moment it's 10 minutes because then it becomes too much of a commitment starting. And then the last rule that we were talking about earlier

is that the data needs to fulfill certain level of complexity. so we don't want to undermine our players' capability, and we also don't want to undermine, you know, the game, the product, and the challenge that we're trying to, you know, solve. So that's obviously the MMO. You've got this huge body of players that are in there. Can the same sort of model be moved to just maybe single-player games

or games that are just not so massive, essentially? Oh, absolutely, absolutely. I mean, the thing is it's all about, you know, putting a mass of people on a complex project. So if everyone just do it, just tiny bit, you know, with, with mass, you actually get insane amount of data. And I think this is, this is definitely a model that can be applied to, you know, single-play games.

It can be applied to, I mean, it can be applied to Facebook. It could be applied to anywhere, basically, or anything. What's next with this? What's going to be in 10 years' time? What are we going to be doing with this? this sort of data? Are we going to be finding more than just exoplanets? How much can we expect the project discovery

to be really sort of discovering and helping our lives? So we definitely want to, I mean, right now, we're working with data from the Korot telescope, and we are looking at all the telescopes as well, and possibly data from further away. So we want to keep the exoplanet project going for a while. But I mean, hopefully 10 years from now, you know, this will be actually doing citizen science should just be, or crowdsource science, you should just, that should be all over the place. That should be, you know, a thing that everyone does all the time.

Well, that's my hope at least. And, you know, it should be one of those just normal things, you know, when you're driving in your autonomous car and it's super boring. You're just, you know, you're helping out solving the world's problems, you know, one tiny piece of the time. That's Berger Finn Borgerson talking about Project Discovery. You can find out more and start your own search for exoplanets by visiting eveonline.com forward slash discovery. In the April issue of BBC Focus magazine, we continue the search for Exoplanets by taking a look at Project Blue. This audacious plan has a single goal in mind to photograph an exoplanet in the habitable zones of the nearest sun-like stars in search of a potentially habitable planet.

You can pick up a copy of a copy of. the new magazine from the 4th of April, where inside you'll find the lowdown on Project Blue and much more, including real-life robocops, how freezing patients can save their lives, and whether geoengineering could cause a climate war. Thank you for listening to the Science Focus podcast from the BBC Focus magazine team. We're 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. 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. 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. First Citizens Bank. You can't reason with the sun. Trust us. We've tried. This summer, it's time to put that angry ball of fire on mute. Columbia's Omnyshade technology

is engineered to protect you from the sun's harsh rays that can burn and damage your skin. The sun is relentless, but so is our gear. Level up your summer at, Columbia.com to spend more time outside and less time slathering on allotion. You're welcome. Columbia, engineered for whatever.