 Welcome again to everyone in the audience. Welcome also to everyone watching via livestream from around the world. We have the pleasure to welcome Christoph Engelmann now. Christoph is an author, a teacher, and a researcher. He researches at the Institute for Advanced Studies on Media Culture of Computer Simulation. That is at the Leuphana University in Lüneburg. In his talk today, he will focus on the enabling dimensions of drone warfare. That's a little spin. By especially looking at data gathering and the mapping of social graphs. We will have a chance for a QA in the end. So please take note. OK. We'll come back to the translation here, if you hear us. I hope we can help you, although we have a bad view. Give us a feedback, hashtag C3T, Twitter account C3lingual. And you hear today Sebastian and ScriptKiddy. And here we go. Thank you. Thank you. Thank you. It's exciting to be back. It's exciting to be back here. This is my third talk at the congress. And as you may know, I talked about history last year. And how you can see through the technology of identification through media. You will see if you watch the video. What you see when you watch the video. I slept this night better. I had a little coffee, but I definitely need more. But what I'm going to do today is to ask a very similar question. As I have often said in the context of drone warfare. And this time in the train of drone wars. When the debate about drones came to light, in the last 10 years, I was wondering, how do you identify the people who see you in your videos? That's a huge part of the war against terror. And how you discriminate against people, who you want to kill and who you don't want to kill. And drone kills people without discrimination. That's... But at the same time, we can see, for example, in the interceptor in October, that there's a very bureaucratic process. Being put in place, at least in the American context, for the Americans to decide the whole thing. So, this is meant to be as your tracker, designed to identify the people you want to follow. There are a lot of people involved. There are a lot of hours of surveillance. And there's the idea of an eye that never looks away. That you look at people every week. And that you... And that you start from the beginning as soon as you focus on the target. But of course, it's a bit suspicious. Because the quality might not be so good. And I couldn't imagine that they use biometric data in these video streams. Because it's a bit difficult. So you have to use something else. That's what I'm trying to... Maybe I have a few answers here for you. The whole research is a bit difficult, because there's very little information available. You have to look for it from the SNOTE data or from other leaks, or from some other scientific work, from some doctor's work, who are involved in it. But it's interesting sources. You can find a lot of interesting scientific works there. So one little thing that I looked at in 2008 and 2009 in the Bayard. It's the whole thing with RFID, with RFID tags, or with technology that somehow works in the optical spectrum. Medium that then will show up in certain spectrums. Oh, the cat's here, it's a shame. But that, again, this seems pretty difficult to be. Because you have to have access to things that belong to these people. And the question is, how exactly can you discover these people here? So you need, as I said, something different. And to continue this argument, that I'm doing today, as I said, if I'm talking more about the drone and its infrastructure, I have to take you on a small road or, as I said, into the history of graph theory. And the ones you know from the computer science or sociology, I probably have a lot more to learn from graph theory than I do. And I am always looking forward to new insights. Here is a piece of paper, which has already been made. And it's about the background. a river, the river in the Swiss Mathematica Leonhard Euler laid the foundations in which this city and this river was located here. He was in the St. Petersburg Academy and published the so-called Königsberger-Brücken problem. The question was, can you go to the seven bridges in Königsberg without using the same bridge twice? There is a way. And Euler's solution was to abstract this geographical setting in a setting that you can imagine like this. These two graphs are isomorphic. They look different, but they contain the same similarities, the same topology, the same topology and therefore are the same. And what we see here are four points, the so-called knots and various connections that represent the bridges in this problem, which are called edges. The knots are the mass of the land and the edges are the bridges. And Euler could show that you have an uneven number of bridges in Königsberg and that this makes it impossible to cross every bridge just once. A graph is usually written like this as a number of edges, as of knots and edges. And the more edges you have, the more weights you have in a given graph. The number of edges is often called the weight. What I showed here, it is important to remember that this is a visueless tool, a visueless tool, but more a mathematical description with which you can calculate. Graphs have found a lot of applications in many areas of chemistry, such as the relationship between atomic molecules in linguistics or new sciences where the connections between neurons are described through graphs. And some of you may have seen this presentation because it comes from the NSA. There are two photo sets that are publicly available, a nice journal for the public consumers of their website. And in 2014, in the volume of where this graph description is published, this is what is called in the language of the research connectors. That is a graph of the connections between the nerves of a given brain. That is a pretty big problem. And last but not the most important, last but not least, we have here, of course, also data structures such as data systems, an application for graphs, for dependency matrices, representations in compilers, and so on. Another more important evidence, which was very influential by graph theory and the most important one for the war against terror, is sociology. And let me just give you a very brief insight into this strange way in which this debate was taken in sociology. So they didn't enter psychology or not. The topic didn't go directly into sociology. How often, in the new paradigm, it came from the side. A very well-known person who is actually known for psychotrauma is... You should maybe keep an eye on this at the conference, and, of course, about the social form in the 1920s and 1930s. They used this whole thing and they used psychotrauma to use the whole thing for their customers. And what Moreno did was set as a spring line. And one of his psychological geographies, a book by him, written in 1934, is a school class in the first year. The medicine on the left side, the boys on the right side, there are a lot of connections between the people in the class and between the girls. And he actually drew all the connections and just asked the people who they were talking to. So who? And after two years, the same geographical arrangement looks like this. And it's practically separated out of two people who go over the gender line. And this already looks like a graph, but it's a couple of mathematicians, but it takes a few trained mathematicians to understand the connections. And if you want to read it further, there's a free book online where you can basically read the history of these developments. So what the mathematician did, Alan Barvelas, worked with psychologists from the 21st century at MIT in the research center for group dynamics. And he had two PhD students who had the first formal definition of what we now would call sociographs. And they invented more or less which we now call what we know as centrality. And mathematically, you can measure how strong a graph is connected to the rest of the graph. And the context in which this research was looked at was companies who had an interest to know the difference between the formal and informal hierarchies in their institutions. So this is a sort of sort of sort of sort of work organization methods where you have very, very strict hierarchies. The central is the information flows. Another important research field in this context was pharmaceutical companies who wanted to know how their medications into the market and how their products and their medications could be brought into the market. How the opinion of doctors could be influenced. So the relations between doctors are characterized to find out which one would aim to address for their campaigns. These were in the 50s and 60s. And this is the context in which centrality measures were first developed which then moved into biometric research to find out which researchers whom, quote, not biometrics, skip a little bit here. Okay, so far for the historical outline. So this is the historical overview to highlight the context in which centrality measures are used in other sciences. And the studies of public and written communication are important in groups in which centrality measures were developed and in which they as an index for stability or instability, stability of organizations were used. And what we call social networks led to many new mathematical concepts even before computer came into the picture. So now in the year 2000 moving forward, we see the graph industrial context. Google first proved that page rank and web search is at least a graph theory question. Also research from the 50s with the context I just described. Google proved that these measures are essential to measure the unmeasurable dynamics of the internet and showed that web pages are modeled as nodes and links between them as edges so that you can really make sense of it. And since then there are new agents new actors and the relationships between individuals and their actions as graphs. In the year 2012 the Bartling's Film & Gardener offered the following point of view. They say there are five essential strategic graphs in the economy. This is the paper. It is paid content if anyone can call me the complete paper I would be grateful, I only have parts. This is a part of this report. The five strategic graphs social graphs search graph consumer graph interest graph and mobile graph. Interest or interest is not quite clear, but this is what Google the consumer graph is a moment of Amazon the mobile graph is distributed by several telecommunications companies and the social graph is more or less Facebook and Twitter at least in the context of western industrial nations. Most of you probably already have the connection. It is just the economy that the importance of strategic graphs does not recognize. According to the Snowden statements, the German Bundestag has called a research institute to the extent of surveillance especially in Germany and a part of the process was an invitation for the NSA whistleblower William Binney who was also here at the congress I think two years ago and asked by the members of the parliament about the importance of strategic graphs he answered we build a relationship we build a social network we build a world there is a document on WikiLeaks 186 186 pages in which it is explained how the world was converted into a graph social graphs will be 15-16 times mentioned no one asks what it is or they ignore the topic interesting is if you read the paper they ask a lot of questions about career within the NSA how to achieve in a position like him how much money he earns he says it is 20% less than a US senator the document is a protocol and they are interested in how far Germany is and the other topic is that the NSA at least in his words is the source code for this project in the early 2000s Binney's project he is Thin Thread we do not know exactly what it is but it seems to be a huge big graph database graph database and the difference that came out later when he was an NSA whistleblower was that basically he did the second thing and it was not a complete they did not save the entire internet traffic and they did not solve the data traffic of US citizens if you and Binney Thomas Drake the other whistleblower if you can believe them is the successor of Thin Thread a project called Trailblazer that was implemented by private contracts and that this restriction of Thin Thread no longer exists so this is the problem it is astonishing how strongly they how strongly one speaks against Trailblazer when Thin Thread basically does the same the whole world and another interesting element is the timing Binney says in the NSA investigation that in the early 90s this year the project was started and he then worked in the 90s and if someone knows the necessary resources for graph data processing so what these companies did key value stores, map reduce and other requirements then you can you will agree with me that it is pretty amazing that it should have been possible in the second half of the 90s so the important part here is not only about companies but also governments who understood the meaning of graphs and how to use them and what happened in the last 15 years between about 2000 and now there is a continuous race to put the world in a graph there is a military academic graph industry complex so we have here the so called stacks Google and Facebook and if you look at the western world if there is less we see that China has different companies in Russia there are V-contact so these are all invitations to you to be transferred to a graph all these systems are very happy would be very happy to have you as a node on the other side we have NSA and the Five Eyes countries who deal with it to graph the whole world and if the Snowden Papers only contains information about these things because they focus more on the access paths and the collection of data basically all the acronyms we know basically deal with the data and less at least on the basis of this paper with the collection and analysis and 2014 we had an explanation since red and these are the graphic tools that are used to find the targets for the NSA and these are for example the photos that are available on the internet it's very interesting to read it's a practical question and answer how it should work we don't know if the photos are real but what is practical is that you have a graph so you graph too much and then you use intelligent so espionage tools to generate more information in the green blue here are the which you want to follow and where profiles start to be applied so what do you do here if you have for example a graph of a terror network that's a question with a lot of debate around also in the missile-based field also the military-based work these are so-called black network these are the evil and the interesting thing about thinking there are many people and I'll show you how they think the report that was published in 2006 in I don't know how many Americans are here even the Times that's a report of the U.S. military and the doctrine about attacks and depending on how the war is going I didn't want to start but 3-4 years in the war they found that we needed a guideline so they have this type that became the CIA director David Petraeus and the publication of the book that's a very interesting book that you should definitely read even if you're a normal author but it's not a good author but it's still a military expert but if only half of the stories that he tells are true then it's really fascinating and he writes the book and of course there's a reference and how you can use social networks and here's an example of how they imagined it so basically what you see is a what you see here that you see a terror network or whatever you want to call it a group and now the important part is what you should look at is the form and what you want to do is to form the graph of the enemy so in the game and you have the graph on the left then you do and then you get a few people and make sure that you can show the graph and the idea is to form the graph that information shows who's important in this graph you can I already learned that and there's a lot of debate about how you do it and what's interesting is this scientific work and in the network center there the and in this work they use the Tanzania data set and this comes from 1989 from the U.S. Embassy 1998 attack on the U.S. Embassy in Tanzania and what you try to do is to automate how these decisions are made and the idea is to make the graph thinner by making it more more central so what and the algorithm tells me which nodes have to be attacked by a network and so the quality here is a bit bad but basically the you don't know who's reloading and the algorithm tells you this point, this point, this point and then I feel the communication to reload which seeks to find a set of nodes who have to be attacked by a network more at centrality and the algorithm will take those people out which centralize the centralization and this is really important in certain networks of often includes restrictions against certain individuals and there's also even hope that doesn't happen that you increase the centrality and which points have to be removed and this is basically an algorithm that helps how to form a graph so how do we get these how do we get the graph when the people are not always online and don't have them around in a country like Afghanistan what did the military do there? they have anthropologists asked these questions which 5 people here do you know the longest which 5 people here with which to last contact which are the most important people here and which are the most important people here and this was converted into a graph the population in this area was converted into a graph and then it could be formed so this is map human terrain the name of this system was made a year ago from 2006 to 2014 it was a very controversial business the American anthropological society actually was confused against the system it was strongly protested there is an interesting film about it which I can recommend it is not really clear I am not really sure if this software was used from the handbook of the map from a handbook of the system what you find in many papers is a software tool an analyst's notebook with which you can find a master of centrality and use it to follow and to create profiles and this is from Palanti which actually controls the market for this context did someone work on a Palanti workstation I still try to find someone on it and I really recommend to watch the YouTube video about counter-terrorism from Palanti it is really interesting Palanti is a major startup founded in 2006 that saw a meteorite rise started as a software product that created a graph for government positions for banks but also philanthropic engineering useful to society yes, but why engineering this is the analytical level what about what about the drones we have to think about drones less than a tool more than a tool more than a tool to discover the metadata of communication if you look at the presentation on the web you can see all these ports the loads the cameras but it is also about listening to communication networks there is the part about communication networks all communication data are taken against and these data are collected and in a graph we have to think about drones as crawlers just like googlebots websites as crawlers in an area to crawl and index and to give the opportunity to create a graph there is also the other side they do not just collect data they also take things to the target and nodes of the graph that they have created another point in the drone debate of course because the first versions were the Marshal-Blood body with a very strong very strong explosive ability which caused a lot of collateral damage and since 2005 we have a race in the direction to find ammunition which takes the individual to the target so personal targets I do not know what personal targets and sub-lethal there are many systems in the competition this is called wiper strike we have presentations about the explosive areas of different weapons and this is small compared to a herfire missile this is what shaping can do there are two ways to form a graph genetically and not genetically genetically is the euphemism for the killing of people if we look at it in a doctrine context in the history of the air we have a movement of weapons to individual weapons of destruction in the last we went to Iraq to find mass weapons and at the end individual weapons of destruction we do not know what happened in Iraq and what comes out in the end this is ongoing research I present basically my tool box has to be filled out since 2000 or a bit earlier a bit longer is the graphs are geostrategic values there are between different institutions a race to put the whole world in a graph and the question who has the largest graph is important and the difference between map and territory is disappearing it is a dynamic representation of activities and communication of individuals and this is the moment when you observe it and there is a different level of usable intelligence it is also very difficult to get away from the graph communication is still being seen as communication and what Binny and Snowden always highlight is the metadata because this is what graphs do they work they delay metadata to say who is connected in what form how close it is very difficult to stay out of graphs not to appear and to develop tools that actually give wrong hints to wrap this up we have to think about who has the graph what it means and how we have to deal with it thank you we have 20 questions for questions if you have questions please go to one of the four microphones the aisles we are going to start with a question from the internet while everyone else is preparing for the microphone one more question from the internet could you tell us how the YouTube movie about human terrain graphics software is called one is the Palantir demo with counterterrorism and I forgot the name of Human Terrain just google I don't know if it is on YouTube left side first of all thank you one question about the creation of the graphs if you create such graphs then of course you have measure errors it is very difficult to create this graph most of the quality measures are sensitive to measure errors my question do you know any research about how new these graphs are and the problems if you decide on the basis of this graph do you kill the wrong people I I am not familiar with the literature but this is a question that beanie is asked but it is not answered in the NSA research the answer is the right if you look at the MIT network one of the things you will see is one of the things you will see in the drone paper what we try is to automate the decision and the answer to the machine or the answer to as many people as possible there are several several processes that are being reviewed to decide what the goal is how I think you will try to deal with it but I am not a technical expert for this question over here thank you very much I just wanted to make sure that I understood the shaping of the graphs means that people are killed just because they are a knot in this graph that makes the graph more central not because they have done something bad or because of terrorism just because of the graph which is called shaping so I don't know I probably without without safety to get this information what I have shown is really what is being set up but the instructions tell practically that when for those who have a certain influence and there is the possibility to kill them or to get information out of them. Thank you. Then we have a question on this side, on the microphone in the front. Two questions for you. First of all, thank you for telling us about the American perspective. It is very fascinating to see how strong the European context is. Now there are two poll selectors off the ground. Have you looked at the application of biometrics onto these technologies? And then the second question is about prediction. The prediction technology with similar... There are prediction skills on other platforms. Have you looked at how they are used to take individuals into the target? I don't have the part with the UAVs to say the technology is used to get selectors. No, I haven't. I haven't looked at the technical side to give you an idea of what I'm doing. I'm interested in how to change the media. Or how we were built from a state of practice on paper. And how we move the new time and how the new tools change the landscape. How to deal with individual identity. So I'm looking at that level. I'm not an expert in technical details. But look into the snow document. There are a few catalogues where there is information on how they collected the data. I don't know anything about biometrics in relation to UAVs. And I don't think they will be used because biometric identification has too many wrong results. And if you are able to identify people. Hi, I'm Henrik. I'm Henrik. I'm a journalist. I've written some stories about the snow document. I can understand the frustration. If you can't see much of this material and can't get it. Can you talk a little bit more about secret research? There is always a balance between universities. So the people on the public side can't get the material. Second question. The relationship between the graph work in the commercial area and the work in the world of secret services. Is one research far more than the other? Well, first of all, the top secret research is what you have to accept. That I work with just leaks and publications that come from the military. And one way to deal with it is basically to deal with the history. So you just go back in time and try to put together what happened in the 80s and 90s. And you can't just talk about the present time because you don't have the information. I don't know what I think it's going to be, if you look at the document. But obviously when you look at the documents about academic debates in the early 2000s, that always goes down. And there's no confusion. But the debate where people from the military talk to other people, and then they talk to the NSA, and then we have a secret service where these ideas are floating around. Other people made a business out of it. Other people made a business out of it for users and the others for governments or things like that. So I think that's safe. I have no idea if it's good or bad. The GARF possibilities of the GHQs or the NSA are there. But I'm pretty sure that they use Palantir. Palantir is that you can deploy fairly quickly, and you can get results with it. So there's a commercial sector that is very strong to decide. Hello. Thank you for the talk. One more remark as a question. You're asking if biometric identification could be useful. One idea that might work is movement patterns. There's a lot of research about it, and it even works on the basis of people that only are five pixels tall. So what you should find is how you can use life patterns. Practically, with whom you stay and spend time, and for example where you sleep, and where you are, and part of this target identification is practically this pattern to find out in the video, and that's something that's used for your freedom. And then you can use it for your freedom, for example, where the person will be tomorrow. And that's definitely part of this process. But other biometric processes, such as facial recognition, are not used. There's a different field, which means against identification. They use retina scans on the ground, not out of the air. I don't think it's available for drones, but I have the feeling that running is something that I can come up with. Thank you. The internet has no questions. Okay, then we go back over here to the side please. Please speak directly into the microphone. Sure. I have two quick observations on this. One of them is that there's actually kind of apparently a little open source ecosystem forming, so a few years better. There's a little ecosystem that's formed and that was raised here with a lot of scientists. And then there's an open source database that's used by scientists, but it's actually from the NCL. And I wonder if you've seen that too. That these data banks are used by the NSA, but then you end up with very simple things. And that looks like very simple matching. And, yeah, head in the clouds. And in which direction does the cloud go? The second question, I don't know, because you can't really look into this project. But I would recommend that it's pretty easy, just a question of the scale. Many, many data points are needed here and they have to be stored for a long time and searched through. A person wrote a book called The Blackstack and it's called The Big Fish Train. And the data you get is a strategic resource and others might not have access to it. So what's the accumulation? I thought it was interesting. If you look at Hackaneers and all these other websites, you'll find a lot of discussion about the software infrastructure of the stacks in the big five countries, like Amazon, Microsoft, Facebook, Apple. And I wouldn't wonder if there's a lot of knowledge and tools that will go back to us. A graph that shows how the data from the next generation probably doesn't look like the one from Amazon. Maybe it's even part of Amazon. So it's part of the big complex. There's knowledge that's flying around and that's in the academic or business context. This course that emerged once so much later was the same. Thank you. And then we have one more question on the left side. You have a question on the left. You told us about the formation of social graphs. Is it the only operation of social graphs? Is it the only operation of social graphs? Is it what's being done? Or is there research on how to integrate nodes? There's nothing about nonprofits, there's nothing about to give anything, there's nothing about appropriation. There's nothing about what people say when you share them with someone. The purpose of advertising is shaping. When Facebook gives you a timeline and then it clicks on it, then it has successfully shaped it. In a context of military doctrine, of course, there is the possibility to remove knots, but the knowledge goes far beyond that. It goes far beyond that, then just removing a knot. Okay, a question from here. Hi. It's not really a question, just a few thoughts about the previous question. Qimolo is, in fact, an open source project by Age Space, and I completely agree that it is in a certain sense different, but ultimately the core is the same technology. And recently I noticed that GCHQ has published an open source graph for distributed graphs, what is it called? I don't know, I don't have it on my phone, sorry. But what I wanted to say is that the industry has really similar tools like what the others use, I think. Thank you. Thank you. Are there any more questions? Then please go to the microphone now. Internet is silent. Okay, thank you very much for this talk. Thank you for listening.