 Hi, and a lovely good evening. The second evening of Divock Reboot to Respond. This is the last talk for tonight and it will be about art. It will be the ZLDKM, they have been in chaos for many, many years. They started at the 19 C3 where they got to know each other and now they do projects together. Events, museums, companies, trade for halls, public spaces and on the internet. They have one name but they are quite different. They have both got the university qualifications from school. But one of them had it through evening classes. They have both graduated, they have both published scientific papers, one studied art and one... They were both born in the divided Germany, one in the Far East, one in the Far West. Both entrepreneurs, one in the programs, they are from Aachen in Hamburg. The old audio suddenly cut off. Welcome in this internet. We are happy to be able to broadcast to you here and become a part of the digitally distributed online chaos. The motto, Reborn to Respond, restart to reappear, but it's very fitting. Not just in the Easter context which has something with being reborn of course or rising again. But it fits our project quite well because it was quite the same there. We said, yes, we have to restart, we have to regain visibility and do something again. The project is quite a child of the pandemic. I think without it, it wouldn't have been created. We are ZLDKM and behind this rather unwieldy name, there is an abbreviation. It means two people that do art. The German words for that spell out ZLDKM, that was a working title. And as it often happens, the title remains. So that's what we are. So our focus is on the future and we'll start straight away with the first question, which is why. It used to be the case that we would meet in a place, we would sit at a desk and work in the same workshop, work on the same thing. That is impossible, has been for a while now, or yes, not so well possible or impossible. So what do artists do in a situation like this? We said interventions, interventions in the sense of, yeah, it's all very unpleasant right now, but yeah. We say waiting is not an option. We need to move forward and do a few things. So art as an experiment and the question to be researched is, can art be done in this way? Is it possible with the given restrictions to create art, to be creative and to develop new ideas and innovations? And if so, how? Our focus, just to make this transparent to you, which is why you see this frame around these bullet points here. We're in this frame, but there is an outside as well. So just to make this clear, our aim is to have art as an immediate result. That's what we define. There is the opposite as well. You could say that the process of making art is open, but we decided not to work in this way. We said there has to be a result at the end. The second item is less who, more how and what. In an artistic context, it's often important who does something and less how it is done or what is being done. And we said if we have a scale of balance and if we decide what is more important than the other, one example is this talk here. It is important to us to make a statement here, to create an approach rather than saying artists don't speak, they don't explain. We are not going to say that this is bad or good or better than other approaches. It's just as valid in an artistic process to say, well, we'll leave it to the observer to interpret. And we are not going to make any statements, but we decided to do it differently. We will make a statement. We will explain art as a motor of change, as inspiration and as a mirror. These are meta levels that we could think about after the talk. We believe that there is quite something to find out if you want to recognize. You are welcome to think about the term of art. You could say that, well, this is a technical, just a play with technology. Is it art? How far does it go? Is it a technical realization or something? Is it from at what point does it become art? We cannot answer this question. We can only invite you to think about this. We say art has an immediate, intermediate aid. It's not a hidden credit that we have to pay in the future. It is an investment in the future in which we want to live in. And we mean that quite seriously without a pinch of salt because reality does seem to be lacking a course right now. We started in February, so this project is about two months old. Just to give you an idea, we used Big Blue Button as a think tank. And we added a wiki for documentation. And that's where we can start. So what did we actually do? We looked at digital art and net art. We ran a market analysis and researched. And what we can tell you is what is out there is incredibly large. There are so many things and it's fun to look into them and immerse yourself into them. It's a bit like reading Wikipedia. You start at one page and suddenly you have 50 tabs open. And it's much the same with net art. You can come from one to the other. It's a lot of fun. I can only recommend that you try. And it's a good alternative to a streaming evening. In our market analysis, we identified a few large topics. We found several, but there are three large ones that we would like to point out. For one, there is the problem of access. If digital art is to be consumed, you have to overcome a certain threshold. It can be a small or a large one. In the easiest case, it's just using the web browser and accessing the internet. But it goes on to having to install and use certain apps, which is a certain hurdle. If you are limited to certain operating systems or devices and have to fulfill certain technical requirements, if you have to have certain frameworks installed on your computer first, if you have to use certain hardware such as gaming consoles or have certain games installed on them. And sometimes you have to keep to certain times. There are exhibitions that are in games on a game console, which you have to have in the first place and buy. You have to have these games. And then within those games, you have to reach a certain competency or level. And often then you have to email the artist perhaps and arrange a date. So that makes it all kind of difficult. It's much like having to make an appointment at an authority in a different time zone. But if you manage to be in the right place at the right time, it can be worth it. The reasons why exhibitions are not available online indefinitely has good reasons. One keyword there is vandalism. Art often is conditioned by technology. It means that you have to use a certain technology and if you use it in the way it is intended. The question is, is that now art? Those of you that come from the surroundings of the KS Computer Club will know Vau Holland. And you will know a few quotes of his, which is one of it, which is a hacker that tries to find a path. Is someone who tries to find a way of making toast with a coffee machine or soup. So the thinking is, you have a device, you have a technology. And you try to use it for a different purpose in a sensible way, in a way that creates sense, but in a way that wasn't intended in the first place. So if you just look at this first item, too easy, not very exciting. That is the way that art is normally linked to technology. So you imagine a smartphone with an augmented reality app and use that to walk through your apartment. To perhaps look at pictures on the wall that aren't there in the first place and swap them for others. And maybe have an artist there that you couldn't normally buy because it's too expensive or too available. Well, that is fun. And it may be an excellent use of technology. But the question is, is that now art? And the question is, what should be the level of the creation? And on the other end of the scale, the topic is, is it too complicated? Is it impossible to understand because too many technological obstacles have to be overcome? And even for us as technically versatile people, it sometimes was very difficult to get access to certain artistic creations. You had to do a lot of research and experimenting before you actually arrived where you wanted to go. And you even then you weren't quite sure if it if you what you saw was what was intended. Because maybe you had a different firmware, a different version of something that could be side effects. So you're wondering, is what I'm seeing an artifact or is it actually part of the artistic creation or process? So that's the kind of scale. On which in between, which there wasn't that much to be found. Third item, vandalism, bullies and trolls. Some of you may have heard even established exhibitions on the internet were robbed. There was vandalism. There was rioting. So this kind of scene that normally has its protected spaces takes its exhibitions and puts it into this shark pool. And suddenly the whole internet comes there and does things. And that will be some result of a digital exhibition sometimes. But then there are groups that look for things like this and actively assault visitors. Destroy things, maybe record this, harass people and then are happy with that and celebrate their success. If they were able to scare people away. So you have to think about these things. And what did we deduce from that? First, it was important to us if we do art in a public space, you have to have easy access. It should be possible to interact and participate. It should be interesting, even for people that don't understand how it works. But it should also be possible to find out how it works. It should be resilient against vandalism. It should be a bridge into real life, through interaction perhaps. And we want to make it possible to create this connection to the world of the art and galleries. To not be restricted in the digital bubble that we probably know fairly well. But we do want to get into the real world. And those were our deductions that we had as objectives. So what did we start with? Our first topic was more or less randomly chosen. We started with Piet Mondrian, because we liked the epoch, we liked his pictures. One of which you can see in the slide. It's black lines with colors filled in. It's about 100 years old in 1921. He started doing these pictures and he's one of the co-founders of abstract painting. And I personally really like his art. And that was the reason we started getting into his work more. Piet Mondrian was often cited and his style was cited in art, architecture, fashion, advertisement, popular culture, you name it. So his art is very well known visually. And this kind of image is well known. And which is also something that is also fascinating is that Piet Mondrian saw himself as a painter of landscapes. Something you wouldn't think from while just seeing his images. But I can only recommend you to get into his art and his way of painting. Because he started with picturing man standing on a horizontal line. So that is his 90 degree angle, which is the basis for his artwork. And in this 90 degree angle, he sees things like dichotomies like man and woman and other images that are opposed to each other. Below you can see how he put that style into fashion. And what we do is we try to make this format digital. So what do you do when you want to convert something to the digital world? Of course, you need computers. So we decided to have computers generate Piet Mondrian style pictures. So use the computers to make the image. Our automatic Mondrian composes these kind of images. What do they consist of? First, every picture is a rectangle. It has a certain size and a certain format. It can be portrait landscape, portraitic, whatever. And in this case, well, in the Mondrian case, we have divisions with black lines and assigned colors. So we need an algorithm that divides a rectangle and an algorithm that assigns colors. Mondrian used a lot of white, a little blue, a little red and some black. And to do the division, we had two different approaches. The first approach is seems a rectangle divides it and the divided rectangles are then further split up. And we do that for while we do a few runs of that until we have a satisfactory image. The second is you take a horizontal or a vertical line. And to make it more interesting and not too simple, we had a third method, which is inserting a smaller rectangle into the big one, which you see on the right side. And then take the border and divide that into rectangles as well. And then we also take rectangular lines from the borders to divide up the rest. The second method is the exact opposite. We do a maximum division first. So the maximum number of small rectangles starting from the edges. And then in the second step, some of these lines are deleted or merged into rectangles. And this is an example of how this can work. And in the end, you have a Mondrian-like structure. So now we have an algorithm that is able to create images in a structured way and also to assign colors. But then we still have the question, where do the lines go and which colors are assigned? And to do that, the algorithm needs input. And this is our version of digital inspiration. We put a lot of thought into that and we want to present that to you here. Mondrian's inspiration were landscapes, but also his environment, his perception of the world, whatever happened in his brain, his brain chemistry. And I'm sure that there's some influence to his art that we're known to him and others that weren't. And that is what we wanted to adapt and recreate. Because Mondrian isn't alive anymore. Unfortunately, we thought, well, let's try it with this internet as the additional component. And also we would like to have a resilient process that might even be resilient to vandalism. We have technical parameters that we create ourselves that we know, but the input from the internet is random. We can't influence it. So this led us to our first experiment. The metaphor is that the brain has neurons connected via synapses and neurotransmitters transport the signals. If you translate that to the internet, you could say the internet is composed of computers, which are connected by lines and the data are transferred using internet protocols. So you see similarities there. So we said Mondrian's concept can be transformed into software. So we developed a program called Pete, which is able to make these pictures, structures, the lines and fill them with colors. So now we're at the point where we have a software, but we don't have inspiration yet. The software isn't inspired. And that is what Poetskene was created for. Well, Poetskene is the derivation of Poetskene, like trainer, trainee, Poetskene. Anyone who uses a computer or server in the internet know that Poetskene's are available freely and it doesn't take long for the machine connected to the network to be analyzed. And you can imagine it like imagine you're walking through a city and approaching a house and looking at the doorbells and seeing what names do we have here, maybe try pressing one of the doorbells and see if anyone answers. And that is what happens with Poetskene's. Poetskene's regularly scan for available services and check whether results are achieved when trying them out. So Poetskene does just that. It fills and probes the internet for responses, which are typical ports used for emails or remote connections. So just your usual ports. And Poetskene registers their IP address together with a timestamp. And that data is collected during one minute and the data created during that time using Poetskene's is sufficient to create a random component to give to Pete. And Pete creates a painting out of these data. And the outcome is in the style of Mondrian. And it can look like that. That's an example, which you can also see on our website, zldkm.art. Every minute, a new Mondrian image is created. And you can see the parameters and the Poetskene's that led to this picture on the left. What else can inspire our algorithms? If you have a mail survey, you know it. Spam is available everywhere. You just need to have an email address. Spam, you have it. Spam. So it didn't take long for us to receive spam messages in masses. And we might just be the first people happy about spam using these emails to create Mondrian images. On our Instagram channel, we have the same procedure. Every day we have a new Mondrian. You can see the picture and you can see the email that the picture was created based on. So now the question poses itself, aren't we creating them randomly and then just randomly assign any emails or data sets and just claiming that the picture is the outcome? What if this is all a fake? And Stefan will tell you why it isn't. Yeah, we want to make it understandable that this is not a fake and show you that these images have a connection to text, which is why we invented what we call the generating report, the generation report. Now, to every text and image, this report explains how it was created and how the text served as an inspiration for the image. And the objectives are quite clear. It is to understand the connection between the characters in the text and the elements in the image and to make it easily understandable. And the design decisions with this quote, random generator, were not to be to run a perfect and wild hashing as you would get in dev random on the Linux, but to do without part of randomness to make it understandable what is actually happening. Now, to make this more understandable, we have an example for you here. Normally, a generation report is a very long text file about 100 kilobytes and it lists all the steps that were taken to create the image from the text, which characters were used, how were they processed, which decisions were made by the algorithm, which values and positions, which rectangles, colors were chosen. And we can't, of course, show you the whole generation of an image, but we can take four decisions and see how they are reported. So as an example, we have your first spam email that we input into our generator and we have pairs of letters highlighted. And what we did with these eight letters all together, I'm going to explain now. The first two letters are HA from HALO. And the first decision that the algorithm has to make, the automatic Mondrian, is how large should the picture be in terms of width and height. And the first decision is the width. And we said that we wanted to have a width between 300 and 800 pixels. And the letters are now used as random input as a dice that is being cast to select a number between 300 and 800. And you can imagine the letter H, which is the eighth letter of the alphabet. So you are kind of throwing a 26-faced dice and you are getting an eight as a result. So you have 26 options for a number between 300 and 800. That's not quite enough. The experts will say there's not enough entropy in this year. So that's why our automatic Mondrian goes and takes a second letter, which is the A. A is the first letter of the alphabet. So we first had an eight and then a one on this 26-faced dice. And using the formula that you see there, we don't want to go into the mathematical details. That gives you the number 209 out of a range from 300 to 800. And that is then linearly scaled to the target. So first it was a random number from 27 to 702. And that is scaled onto the 300 to 800 range, giving you a value of 435. And that is going to be the width of the picture. And the same way we deal with the height. And then we arrive at a point where we have to divide a rectangle. And the algorithm decided that it wanted to use the third subvariant of the first variant. So have a large rectangle with a smaller rectangle in it. And one question that has to be asked at this point, at which horizontal position should this new edge of the new rectangle turn up? And that's what this arrow is showing in that image there. So we need a distance from the left margin. And that's where we take the letters I and R from the spam email. Again, we throw a dice twice and reach the number 249. And that is the distance that we use from the left margin to make this division. And then the algorithm has to decide which color this new rectangle should have. And of course, and for that, it only goes through all the rectangles. And for the rectangle number two, it takes the letters T and I from investment plan, the German word for that, investitunes plan. Again, the dice is cast yielding the number 29. And so that means if you look at the ranges that are shown in the report there, you will see that 29 is part of the yellow range. The ranges are not equal weight. The next rectangle we'll show you is number 20. And the letters E and I from the spam email yield white and so on. And you then see the picture that emerges with the rectangle 20 being white and rectangle two being yellow. Now that in principle finishes the generation of the image. But the actual painting of the image, of course, is only the first step, the main step admittedly. But after that, the image continues to live and normally goes all the way to a gallery. Yes, and we've said that it would be one of our objectives to create a bridge to the world of art and galleries. And for that, we have, we saw two options. We found two options. The first one is that we use the electronic virtual product, the image in PNG format, say. And enter that into the art market. But that is probably quite difficult because the market for that is fairly small and it's not easy to communicate where the creation is in that, why should this be bought or stored as a work of art, put up on a website. The problem is that you can't really recognize originals that way. You would have to introduce some copy protection mechanism, which we don't like that much. So that is a difficult thing to try and do. So we thought we have to find another way. And another way is that we take these digital images and take them back into the real world. And for example, create photographic prints or other prints through different technologies on canvas or acrylic glass or something. There are different options there. Connect it with the report. You can sign that. And thus, you have an approach to entering the gallery market. There are originals that you can touch physically, that is tangible. And it's clear that there is an original there. And we have another variable there, which is inspiration. That can also be used in the arts market context. You could say that you link up with authors of poems, for example, and say we are going to have 10 poems leading to 10 images. And maybe create a sort of synergy from that. As an input channel, you could use music too or stock market rates or whatever. And one example of such an image you can see here in this picture, in this photo, which kindly was made available to us by Sascha, who bought our first picture. He actually put it in a frame, creating the first original work of art. But then the question is, how can the price for such a work of art be determined? So first, we looked at how galleries in the analog world determine prices. And many galleries use a very quite simple formula. They use the size of the image, which is width plus height in centimeters, and multiply it by a so-called artist factor. Now, this artist factor is in principle a measure for the fame of that artist. And as a guideline, you can say for beginners, you can say three euros per centimeter. And what we now have to do is kind of transfer that to the automatic Mondrian. And what we've already decided is that width and height should be converted by a factor of 10 pixels being one centimeter. That's how we would carry pixels into the analog world. And what's missing, of course, is the artist factor. And you could say that the automatic Mondrian is new. So it would be a beginner's artist factor. But perhaps the images are quite good already. And well, I'm not going to say equivalent to Mondrian, the real one. But maybe some step off the way towards him. So we ran a test and we asked ourselves, how can we find what normal people would think of this kind of imaging, how they, how much they could tell apart our images from authentic Mondrian images. And because we both studied computer science, of course, the idea quickly came that we should use a Turing test, the test that was invented by Alam Turing to compare computers and real people. In the case of Turing, this was developed to test an artificial intelligence that was to convince people in a conversation that it was not an artificial intelligence, but a human being. And if the human being would believe that, the test would have been passed. Of course, in our case, it's not an artificial intelligence. Things are not as complicated. So we took two images, one automatically produced one and one that was actually created by Mondrian. And we asked normal people that you find on Twitter, of course. Please choose the real Mondrian out of these two images. And you can see in the results that 49% decided on image A, 42 on image B. And 9% said, well, I cheated. I used Google. So we'll take those out of the evaluation. And we can see, well, it's a close result. But the wisdom, swung wisdom worked. Image A is real Mondrian. It's called Trafalgar Square. And Mondrian painted it from 1939 to 1943. And image B is an automatic Mondrian out of a spam email from mid-March. The content of the spam email is a fake dating request, which for reasons of parental protections, we cut here. And we would like to repeat this test with two more images. So you see two images here. One is generated. One is an authentic Mondrian. And with that, we would like to enter into a discussion. So which one of these is real or authentic? And if you have other questions or comments, please feel free to ask them too. OK, that was a very interesting talk, a lot of art and a bit of cyber as well. And we do have a few questions. First, I'd like to thank you for your talk. Thank you to Stefan. One of you two is a data protection commissioner. So the first question now is from someone that perhaps hasn't quite understood the GDPR too well, or maybe has a very unique interpretation of it. Because the question is, if you use IPs and data from a port scanner, what has to be said regarding the GDPR? I added my doubt. But what's the answer? Yes, it's not to be taken so seriously, because these are data that have been kind of pushed down to us, voluntarily given up. So in that regard, you can hardly say that there is a problem here. So regarding the GDPR, I don't think there is anything to consider. But of course, if we do get any queries, we will be happy to answer them. If people would give us IP addresses that at some point might have arrived with us and that might have stored, then we can find those and we will delete them if we have them. No problem at all. And then we have the inevitable question. It cannot be averted in this context. Why don't you sell your art using the non-fungible token? Well, first of all, digital data can still be copied, even if you write it into whatever number of blockchains as a non-fungible token. And to me, an NFT like that, if I understood it completely, it can be compared to going into a museum, a gallery, and buying an image, taking a picture, turning it around, writing on the paper that this is my image, putting the image into a safe, and then putting the image somewhere. And that's not the way we want to do it. It would be an interesting approach. Question in English that I like. Wouldn't the guesses on which Mondrian is real be random? I was just to try with both artificial or both real ones as a baseline. I think that's a very good suggestion. We can try that for the next devog maybe, or maybe even earlier. OK. And then, of course, the question remains, which one is the real Mondrian? C or D? We've seen A or B in the talk. I think A was the authentic one there. And in this case, it's actually D. And if you paid attention, you could have seen it in one of the slides earlier. OK. The vote actually was in favor of B. So you can see how to what extent people are wrong. So there is going to be a session in big blue button after the talk. Or what is it? Yes, it is big blue button. And the big blue button server is fragen.zldkm.art, so fragen is German for questions, .zldkm.art, self-organized Q&A. So ZLDKM, to remind you, is two people that do art in German, the words, the first letters of each word. So if you want to talk to the two Steffans, we have a few minutes left in theory, which unless you want to say something, we could simply skip. But we would like to thank you a lot for coming. I have to trust this talk and take my time for my own question. So thank you very much for being here.