 Feder Bart. So Feder is actually our international speaker. He's from the Netherlands. He's at Delterres and at the University of Twente. And Feder is going to talk about models in virtual environments and digital twins. Yes, thank you for the invitation. Can you see my screen now? Well, yes, we can. Yeah, so I wanted to talk a bit to you today about digital twins. So I work at Delterres and TU Delft. Actually, I'm also organizing the workshop with Robert Jan from the TU Twente. And a bit about myself. So I've got a PhD in coastal engineering. I work a lot on visualizations and models at Delterres. Also once escaped from Alcatraz there on the top right. And I discovered an island, a blatant copy of the Chinese from our Palm Island that we built. Also working on some other topics at Delterres. But data science and future modeling are my main topics. And today I want to talk about one of these, let's say, future modeling topics. And that's digital twins. And so that's actually what we call emerging technology, something that is now becoming popular. And some of you might already have heard of this. And especially in Europe, it's now becoming quite popular. And I wanted to discuss a few things. What is a digital twin? A bit about the European initiative, the digital twin earth. And also what makes it different from other things that we already know. And I want to talk to you a bit about why digital twins are popular now in the discussions with you. And so a bit about the digital twin. And so it's actually a theme or let's say technology that come over from our industry, from building cars, for example, this one of the earlier pictures that I could find. And then people make a virtual copy of a real world object. And they do that so that they can interact with it. So that they can design it. And that's now a topic that also transformed into our environment. Here you see an example of the show. They have a digital twin of planes, for example, and French company. And also in the Netherlands, we see it used quite a lot. And this is one of the tunnels that we are working on in the Netherlands. And then we make a digital copy of that tunnel so that we can test all the things to see if people react to it well when they arrive into it virtually. And that we can also test all the equipment and the wiring that goes in before we build it. And this is now also something that transformed to the environmental modeling. This is an example of something that is now called a digital twin. It was already used to exist. But it's an integration platform of Tiger and in this case, the Dutch company, where they build the cities and also include interactive models, for example. And so and that is a thing that we now see more and more often. Also, if we look at trends, for example, you see the Google search trend of how many people are searching for digital twins actually looked up that the United States is not yet on a high in the list of people that search for it. But if I look at Europe, then people and then it's becoming quite trendy. That also gives an expectation of how this curve will continue or prediction and that we now get to a peak of inflated expectations and then trough of disillusionment. So that's what comes after this. But up to now we have an innovation trigger and now we're moving to this peak. And so if you talk about the digital twin, it actually looks quite a lot like things we already know and we already been using for quite a long time. And so, for example, it looks quite a bit like simulation so far. But then typically it's a bit different, for example, simulation like here Delve 3D model that we are working on. And it's typically passive. So you can start it, you can run it, but you typically don't interact with it while it's running and it doesn't create, let's say, an immersive environment that creates charts and maps with these jet color schemes. But it's not, let's say, realistic. And you also typically do it on your own. So that's what makes it different from a digital twin. There's something else like a simulator, for example, your flight simulator. That's also not quite a digital twin because it's more single purpose. You can't use it for a helicopter, for example. And it's also quite solitary. You sit there on your own. It is very interactive. And it's also what also makes it different. It's not real data. So it's typically sometimes contains some old scenarios, but it's typically fabricated data used for training purposes. And it's also what you see in, for example, series games. Another thing that also has some similarities with digital twins that typically doesn't contain, let's say, real or let's say, well, realistic physics. And also typically doesn't contain real-world or actual or current data. So those are some things which are almost a digital twin. But if you really want to make a digital twin, at least that is what we found. If we ask people, what do you expect from a digital twin? When we call something a digital twin, then people think it's going to be this. They think it's going to be photorealistic, interactive. They think they can touch it. They think it has real data and real physics. And now I'm going over all these things to see what we already have as components that are actually going in this direction. But if you zoom out a bit to see what people are actually doing when they call something a digital twin, one thing that is really remarkable is that everybody makes the same picture. So before they start building a real digital twin, everybody's making a picture with dark background and light blue lines. So somehow that's also very representative for a digital twin. In Europe, we actually have a high ambition. You could say a bit too high ambition. And that's called the project destination Earth. And that is aimed to make a digital twin of the Earth. And when I first heard about it, I thought, well, let's keep it simple. But still, this is now, I think, about several hundreds of millions. This is the direction also from the Green Deal and the European Digital Strategy. There's a nice paper in Science about this. And so you can read more about this. But this is the direction we're going. And they actually also made a nice picture with the dark blue background and some lights. And so that fits in quite nice. One of the topics that we are actually already contributing in this context is the precursor digital twin of the ocean. And that is one of the things where we put in our most realistic models, for example, our global title model. And so that's actually where we're already actively working towards. And so back to the list of examples. And so if you look at real physics, that's actually something we've been doing for quite a few decades. And so here's an example where we have our old flume, where we invite high royalties to really touch the water. This is where I'm coming to back to that. And that we want to let's say that is needed. Many scientists have to really get our tools and models into the hands of the stakeholders. That's also what we use here. This is our latest effort in that direction. Let me just mute it because it has some singing in the background. So that physics are real flume. This is what we use to protect new dykes that we want to build in the Netherlands. And this now can generate the biggest artificial waves in the, in the world. So that is, and we learn, we train our models on these data that we generate with these new flumes. And we also use artificial intelligence to detect all the things that are going on in this flume. And then we transform that into our numerical models. And that physics is now becoming a bit, let's say, grayish area into data. And because these days, as a lot of you probably also do, and we're not just simulating physics, but we're also learning from physics here. This is a deep neural network that we made to learn from to learn a good downscaling method by feeding it continuously high resolution and low resolution images and if you do that long enough. It will start learn if you provide it's more context data is we can generate a high resolution image from a low resolution image. So if you are doing this as well, for example, Google is working on a hydrological model and it's only feeding it with data. And so that is that it sort of becomes a grayish area. Also, if you look at the other way, if we look at the real data. So this is our new global title search model that we use to predict types and currents in the coast all around the world and a very high resolution, but we don't give this model anymore to people we just give the data so that that was sort of becomes a grayish area here you see the search levels that it creates very detailed along the coast, and this is typically one of these components that we put in for example into this digital twin of the oceans and then we provide the storm search and occurrence. This is another example of a data set that we make this also already going ahead like I'm a bit skeptical but if you look at if you zoom out we're actually already working quite a bit towards a digital twin of the earth here you. This was the aqua monitor that we created a few years ago, and there we scan the whole world for new land and water and this is actually where I found this all the way up here I found this little Chinese islands. That was one green dot on the top right there, which was a copy of this island down here in Dubai. This is new land and water and new water here in blue, and you see all these new reservoirs being built. This is actually one of this is examples of the direction of this digital twin earth. We also got a new grant from Google for one million to work towards this global water watch where we scan all the reservoirs in the world to get time series and really help people to act to see what everybody's doing. This is the water from who we're working on that with the World Resource Institute and Wildlife Foundation, and we also zooming in all the way here we're making video maps for example in this context. This is what we do for water peace security and you see that they're new and new them is being built, and that people are keeping the water from downstream and they see people moving here so that is one of the things that we scan for a ministry of foreign affairs. And another part is the social party you really want to focus on decisions. And, and that is something to go with on your own, but to really have people being able to touch the models that we make and to feel them to work on this together and the first example is the three I model where we worked on to really get it into the hands of the stakeholders. Another example with Robert John who's also here organizing the session with is this virtual river where people can really play with the river and designed our own river. And he, Robert John designed this nice, sort of game like environment, and we connected this to this Tigran engine also so you can also 3D view on this monitor. And we connected that to the Delta 3D model to do the real simulation so that kind of things that social aspect is really something that we really consider to really attach it to the real decisions. You also do that in the United States have for example here at a nice skill model of the Mississippi, which sort of a hybrid clone part physical, but it also has 16 projectors that project the possible futures of the Mississippi delta. It's really inspirational I think as a two words the interactivity that it's also something really important here you see the model that you have of the North Sea, where you can really touch the model so you can put in paint. You see the paint being picked up by the model, and it starts to flow as you draw it. And here you see the tide coming from the UK, coming all the way to the Netherlands and this is transporting this paint with the current. Another example I'll go quickly through it is this bond the model, where you can answer the high beat one the 2D model way of channels, and what you can do then is turn on the rain has a can put down rain here really strong rain will then start try to move this water away and but in this case it doesn't succeed so you see these pulled are slowly filling up with water and if you wait long enough it will turn blue, and then you see that it's filled up so they have a problem, and they can start designing interactions as for example here you can raise the capacity of this pump and you can discuss with the people if that would be enough to help them solve their problem. And so this kind of interactivity is actually something that we made possible together with the system said the BMI interface where Eric and others are working on. And so that really helped us to make these models interactive so you can touch them while they're running. And that's also the thing that we think is really important towards these digital twins. And so we'll have a workshop later with Robert john about how to really design this into activities and we also want to get our ideas, because like visualization how to visualize is well defined, but how to interact with models that's a real new playing field so we really interested also in the sharing and gathering ideas with you together. And so, towards the final two things photo realistic, and it's also something that people typically expect from a digital twin. Actually, if you ask them what's important they, they all expect this but they don't think it's important. That's also something that we sometimes do not always. So it is example what we use for the sea level estimates that we try to convert all the radar images into something that looks like snow and ice and transform it into video maps so people can get a real understanding of real feeling that Antarctica is really losing ice at the place that's the most important for the Netherlands. And here are some examples of visualizations that we use to warn people to not drive their car through a tunnel in the case of a flood so we try to show them that that let's say not a good idea. And so those kind of visualizations are now much easier because you have photo realistic rendering a physical based rendering and real time ray tracing so there's really steps going forward that makes it much easier. And so back to the digital twin earth they have the integration. There we see a lot of effort. And for example, from the Pangeo group, but also add in Europe where we're trying to build the integration platform. We're also actively looking at what the cloud providers are doing from that concept so we're looking at the planetary computer for Microsoft for example is using a similar approach to what Pangeo and what Fernando talked about is Jupiter approach that it really gives a new way to work with all these data and integrated. And Google is of course doing the same with the Google Earth Engine that's what we also use a lot. And that's also what you see now that all this data is moving towards the cloud that's almost impossible to keep all these data sets local especially if they start growing in the face that they're doing now so then the cloud somehow becomes the default integration engine and the cloud that is an interesting development, let's say, but back to why are they now popular. And so if I talk to the explain this to my colleagues say usually the first reaction this is just old wine. I didn't say that in English old wine. Just a new something old that we sell as something new, and to a certain extent it is so you also see that in that people start with making the image. But if you look at it it also fills in a real need that for us scientists to really get our models and tools and data and information really into the hands of the policymakers. And if I look back in my archive and once made this 3d visualization and like with real stereo photographic visualizations and it was one of the happiest audience I've ever seen. And so we also really like it ourselves and it sort of fills in. And what we say the digital dream, and that's also what you see with all these dark grounds that is this is this futuristic dream is somehow in need that we have grown and if that need is fulfilled, we're more likely to appreciate it. So that is my view on the development of digital twins and that's where I want to wrap it up. Thank you so much, Vader. That was great and you should have seen the smile on my face I think this was, you would get a happy audience here as well, if it only was in person. So the digital twins give really access to, to what if scenarios right you can, you can play with model simulations or with visual imagery, and you make it available for people that are not specifically familiar with models so I think that that's pretty really see it as a new dissemination format, although it's it's hard to add people do expect a lot more from it if you call it this way as we try to be a bit careful on introducing such a concept. And we have a minute for questions does anybody have a questions if so you can raise your hand to become a new YouTube or you can put your question in the chat. Yeah, thanks, Greg, old wine in new bottles was what I was looking in Dutch it's a shocker so much back somehow. That was from the Bible I think they still had wine in bags back then. So this, we use this BMI edit that allows you to make it to talk to a model, while it is in this in a running state and that is really the essential let's say key part to making something interactive. It also allows you to do many other things and for example, a simulation for the, and what we just saw from Paula, and like a steering a model in the right direction or changing nudging the symmetry that's also some simulating a model is also something that you can use BMI as a basic concept for this also where we, who can our open data simulation platform.