 Thank you everybody for coming. Please welcome our next speaker, Max Rosman. He has a background in chemical engineering and philosophy, and in next talk he will share his ideas with us about how we can make models, which we could use for more analytical processes to make them, how he could make them more immersive through narrative. Yes. Max, please give me a round of applause. So yeah, thanks for this introduction. Well, I'm Max, and as you have already heard, I have no IT background, I'm actually a PhD student in the field of technology assessment. This means we make something like ecobalances and surveys about acceptance and risk perception, for example of stationary batteries, or biogas plants, or the nutrition of the future and so on. And on the other hand, my PhD is about the role of narratives in science. And I actually want to use this talk mainly to come in to talk with you. Well, I was, last year was for the first time at the Gulasch Programme in Nacht. And I left this openness spirit, and some weeks later I started Python programming. And yeah, this was really great, and there were these stupid notebooks, and they helped a lot to get into the material. And then I had some thoughts about term computational narratives, and I wanted to come in to talk with you about so that we can maybe later discuss a little more about the ideas on this. So I will just start with a question. So that's mainly my question on my vision with the computational narratives. And computational narratives improve to the transparency, the collaboration and the participation by connecting the scientific modeling with efficient narratives. So the idea is, you can see on one hand you have something like the public discourse about technology. This is what you read in the newspapers and in the program of political parties. Then social scientists can analyze the discourses and find out what are the main narratives, the key narratives. And on the other hand, there's public participation. And then they bring this somehow into science, and scientists make models to assess these technologies, for example, from an ecological perspective. But if you think of, for example, you have this scenario with biogas plants, and you have another scenario with battery storages, and you want to compare this in the energy grid, your model will be unlimited complex, and this is not possible to model. And if you would try so, no one would understand this model in the end. And so the idea is to find out about the key narratives and then somehow make better models that just follow the key narratives and just take the most important aspects into account so that people in the end can still use these models for their personal assessment. The other feedback circle here is the scientific discourse. So if we talk of scientific knowledge, then it's peer-reviewed knowledge. That means scientists publish papers, and other scientists read them. And if they think this is good science, they will be published. And this is like another world or another system, you can say. Yeah, and the ideas, you could put the computational narratives, the stupid notebooks and so on, and this plays because everyone can read a narrative and follow the storyline, and then you can play inside this narrative and play around with the data. You can look at the sources, you can look at what models are used, and so maybe this can improve this interplay. So the question is, does it serve as a medium or a boundary object for both worlds, the public and the science? Yeah, and well, I didn't give some table of contents, but it's mainly I will look at these computational narratives as games of make-belief, as narratives and as models. And yet, just to give a further overview, here are three quotes that somehow seemed to be related to me, and that's the idea. First thing is from iPython, that's a project page from Jupyter Notebook and so on, Project Jupyter. In order for data and the computations that process and visualize the data to be useful for humans, they must be embedded into a narrative, a computational narrative. On the other hand, this is Roman Trig, he's some philosopher of science who works mainly on this idea. He says, model descriptions are like the text of a novel, they are props in games of make-belief. And on the other hand, this is of the Dungeons and Dragons players book, basic rule books. Like those games of make-belief, Dungeons and Dragons is driven by imagination. And so somehow you could also say it's not only about philosophical reflections on Jupyter Notebook or do coding and role-play games somehow belong together. So I already heard that some of you do role-play games or cosplay and so on. This is some tabletop game from Dungeons and Dragons and about just to say, well, you have got the Dices, you have rule books and on the other hand, you can write your own stories or you buy a storybook and then you're going through adventures, but you cannot do anything, you have to dice and somehow it's then restrained what can happen and what not and there are rules and it's not just arbitrary. So on one hand you have the narratives and this is also part of the Dungeons and Dragons players' basic rules. The Dungeons and Dragons role-play game is about storytelling in words of sorts and sorcery. It has elements with childhood games of make-belief. Like those games, Dungeons and Dragons is driven by imagination. It's about picturing the towering castle beneath the stormy night sky and imagining how a fantasy adventurer might react to the challenges the scene presents. So here you have to immerse and follow the narrative. On the other hand you have the models. The Dungeons and Dragons rule set gives structure to the stories, a way of determining the consequences of the adventurous action. Players roll the dice to resolve whether their attacks hit or miss or whether their adventures can scale a cliff, roll away from the strike of a magical lightning ball or pull off some of the dangerous task. So I also once played this with the Dices. I played mainly on the computer role-play games and yeah, my experience was like this. I played something like some Nomad Trewied and when we entered the first town I wanted to talk with everyone. I wanted to go into the bar and to spend some money for drinks and so on. And then, well, what happens, there was some discussion with the Dungeon Master and he said, well, characters of your kind don't behave like this. And I said, well, but the rule set says nothing especially about this. Maybe I'm not as all the other Nomad Trewieds. They said no, suddenly the sky turned black and some lightning bolt hit the Nomad Trewied. This was like the end of my pen and paper career. But I didn't really understand what I made wrong because I didn't know, yeah, he wasn't in the rule set, what was the problem and, yeah, now I understand a little better what a narrative is about and so we come to the next point. What is a narrative? And here we start with Aristotel. He says, well, he is one of the first one who thought of the language, a link of the power of these tragedies as text and he says, a tragedy is an imitation of action that is serious and has a wholeness in existent. And the important thing is the wholeness in existent. And what is this? Well, it is which has a beginning, a middle and an end. Okay, special, cool. When I read this the first time and this was quoted so many times in this narrative science stuff, I just wondered what's so special about that it has a beginning, a middle and an end. He further says, a beginning is a beginning and there's nothing before, but when there's a beginning, you expect that something comes afterwards. So it's like immerse into a whole narrative world. This is the beginning. So if you start reading something, then there's the beginning and this is like some principle of generation you have to start imagination. So I could say today we learn about narratives and if you follow me with this, then you are immersed. So a narrative is a whole lingual reference to a sequence of events. What I find very interesting, this concept is also used in psychology. And there it is known as the narrative self-constitution view. This is, for example, used for trauma therapies. And it says, narrative self-constitution view says, that we constitute our identities as persons by generating and operating with an autobiographical narrative that meets the articulation and reality constraints. So on one hand, it's again articulation. You have to tell a story about yourself. And on the other hand, it's not only you who tells the story, it's also people who continue the story with you. And this means your individuality is not like you are person X with properties, whatever. You can say, I'm the story of myself and I'm the author of myself and other people are the co-authors and we all together live in stories. So McIntire has some moral philosophy, he says. He also takes this perspective and says, I can only answer the question. What am I to do if I can answer the prior question of what story or stories do I find myself apart? And this gives your connection to your birth, to your home, to all the people that surround you, and to all what you make out of the things that you perceive. So it's about the stories you tell. So in reflection, we are the protagonists of our role-play game, we can say, and we have to find out what our stories or what your stories are about. This is really interesting. This also has an impact to sociology and there we can analyze narratives for collaboration and orientation. So for example, narratives like visions or reports or dystopia, they have an effect on people. For example, if you look at your mobile and then you watch an whole episode of Black Mirror, you will behave differently. Because in these episodes or in the series, you learn what this actually means or what this can mean. So the meaning changes with this. But there are also visions like the car-friendly city, and then people start to build things like FordSign. Or there are other visions like the smart grid vision or the automatization vision. And what all these visions have in common, on one hand, they bring people to orientate what is important today and what will be important tomorrow. For example, one can say, the problem is we are dependent to big industry groups. If something breaks, we must buy expensive spare parts. And no one knows how to repair. That's stupid. Let's start a FabLab and share our knowledge and to become independent. And this is like a vision and if you repeat this vision, this vision can spread all over the planet and FabLabs will rise. So that's the idea how this helps to orientate. On the other hand, there's the discourse of reflection about such a vision. So if you are on the FabLab, you can start, what happens if you are founded by some industry? Are we dependent to them? Are we exploited from them? How much can we be independent if we relate to money of them? And this continues the shaping of the groups or the collective identity. So that the FabLab becomes an individual actor. On the other hand, it's collaboration. So that's the question, what's my role in this game? If you, for example, go to some FabLab, a lot of stuff has to be done. They are hungry, they need some place to be. They need someone who cares about some insurance, who does some public relation and so on. And so it's not necessary that there's some head of this thing. In those collaborative actors, mainly there must be a vision and people can find out what is their role in this vision or in this narrative. So this helps for orientation. For example, when you think of the car-friendly city, there was this guiding image. And while the economists know how to calculate something and the engineers knew how to build roads, streets and the lawyers, they also did something about this and so on. But the most important thing is narratives and visions become active. Like I have a dream and this activates people and so people do something. So if narratives do so much for collaboration and orientation, then there comes Project Jupiter and it's also about narratives. So actually such a computational narrative could be the base for all the stuff that I just mentioned. So that's the whole quote. Computers are good at consuming, producing and processing data. Humans, on the other hand, process the world through narratives. Thus, in order for the data and the computations to process and visualize the data to be useful to humans, they must be embedded into a narrative, a computational narrative that tells a story for a particular audience and context. First, a single computational narrative needs to spend a wide range of contexts and audiences. Second, these computational narratives need to be reproducible. And third, computational narratives are created in collaborations. I will just, for those who don't know, but actually I think you all know Jupiter notebook better than I do. But that's like the structure of it. You can see it's, you open it and you go through it. You follow the linear structure, you start on top reading and then you go down and further down. And here's like the principle of generation. This is some tutorial for machine learning. And you read this and then you can, you know about the context and then expectation rise. So you start to immerse into this game. And then there comes some computational code following certain rules, some visualizations. And then the interesting thing is afterwards, after this visualization, you interpret this visualizations following the plot that was introduced in the beginning. So it's not arbitrary what you see in this plot. It depends to the story. And in this Jupiter notebooks, it's bound together. You have the code and the story and one document. And this is some plot, for example, from my biogas plant. I can tell you now, this is the energy price, the stock market price, and the blue is when it's a good price and there's energy is sold. And the green line is the gas storage. And so, yeah, there's some, like a simulation of it. But the thing is you understand this or you say this is a representation of a biogas plant if you follow the narrative to this point. And then you can ask questions. What is missing from the biogas plant in your imagination? So you think of a biogas plant and you say, okay, there must be some warm storage or there must be some agricultural simulation. Or you don't ask for this, but this depends on the imagination. But how can I distinct a good model from a not valid model and dare the fiction view on model come into play? And again, it's about games of make-belief in philosophy of science. Roman Brick says, model descriptions are like the text of a novel. There are props in games of make-belief. Game-based and public rules are authorized. Games involving ad hoc rules are unauthorized. By definition, a prop is a representation if it is a prop in an authorized game. What does it mean by this? Well, we explore and develop models in the imagination from an internal, our participatory perspective. And we originally compare models and targets from within an extended imagined context. In this case, we are fully immersed in the game and we assume an attitude of imagination towards the relevant propositions. However, we can also exit the game and assume an external and descriptive perspective. So on one hand, we immerse, follow the narrative and we only find the representation in our imagination. On the other hand, we can quit or exit this imagination and analyze the model constraints. And this makes, for example, the electrocardiogram not only a plot of some physical data, this makes it some indicator for the health or vitality of some human that is connected to this. Or on the other hand, this makes this not just a plot with nice colors, it makes this some representation of a biogas plant. But what about things like this? Like a Bitcoin price prediction. Are they based on authorized probes? Well, I found it quite nice last year there was this big hype on this blockchain and coin stuff. And there was a website called TradingView and their people share and discuss their models for Yota prediction and whatever. And it's, yeah, you can look at it as art or you can think or wonder if people spent their money by this and some of my highlight was this. Well, he found out that it's actually a circle and it's sometimes following when the circle goes down, the price goes also down. So he predicted, it made a good prediction for this time because Yota really crashed. And if you follow the data-driven view, you can say a model is good if it only fits the data. But if this is all, you need arguments where this is not a good strategy. This is something I also find quite nice because it looked like this bureau graph things people can paint. Yeah, I actually wonder what's the game of this model. We can say, well, we immerse and follow this and follow the story of you can become rich because you're smarter than the rest. And then there are some props that lead you to immerse and to take this serious and follow them. Or you ask, is this plausible that I can predict the market price with something like this? On the other hand with a fiction view, so we can choose both perspective. We can argue for the model world representation or ask the question, does the narrative refer to public rules? And that's the key of this fiction view on models. If it's not based on public rules, it's not authorized and then it's no representation. So this is especially important when you think of climate prediction. It's not only about does the data of the day fit to the models, it's also about how the model is built and all these public rules that are behind or that play a role in this model. So the fiction view on models, what we can say, the model system represent the target system in imagination on the basis of authorized rules and props. Or you look at the Dungeon and Dragons book. Dungeon and Dragons or the physicists or the VDE rules that give structure to the stories. And this is a way of determining the consequences of the adventurous action. Yeah, and we can choose. On one hand, we immerse and follow the narrative in a fictional model world comparison and reflect plausibility and completeness. On the other hand, we can exit and analyze the logical consistency of authorization, references, and interpretation. And just a link, there's a lot of philosophical debate about truth and fiction and fictional operators and all this is done in logic. And it's quite complicated and interesting or you can also say it's some kind of art depends on what you want to make out of it. So the last two things. This is like my summary then at the end. Data narratives represent the reproducible experience to make within the data model in computational narratives, you as individual and as a collective. So that's the clue of it. Follow a storyline and explore the constraints by playing around with the data, the variables, the model and code. And that's the idea. Maybe computational narrative might theoretically serve as a technical infrastructure for better collaboration and reproducibility. So that's also an important point in science right now. By establishing, organizing, and devolving a shared problem, solving perception with transparent references to the given rules. Now my question is to all those of you who have more practical experiences with these notebooks, do they? So that's the question to you. Thanks. We still have time for some questions. So maybe is there any questions in the audience? So this concept of computational narrative sounds like it would be potentially quite useful for teaching people about technology and about computing. So do you know if there's been any work on that? Well, my experience, when I started to learn Python, I was always referred to these Jupyter notebooks and they helped really a lot. And if I have any question, I'm looking up some Google or something like this and I come to some thing that looks like a Jupyter notebook so far. So I think in this informatics scene, this is quite common, but I also think that it could serve a lot for engineering and for all these other stuff. So I think, yeah, that's mainly a question on you. What do you think? Do you think it could serve for it? I think it could, but I don't know if it's done so much. What do you think? Maybe I would have one thought to that is I feel that narrative is often driven by conflict. This can be conflict between individual actors or conflict between an actor and the world. How do you, maybe that would be a thought, how one could put this kind of conflict or position this conflict in the narrative of a model. That's a good thing. So Mary Morgan was also a philosopher who deals with this fiction-fueling models and she said she compared it with a detective game if you're doing some modeling. It's like you have some question like what is better as a flexibility option biogas plants or battery storages? And that's like the point. And then you have to find out and you are looking up things to solve this question. You look up how a battery is built, how this is built and you're collecting data and comparing data and you're following this and this is also part of the structuralist narrative discourse analysis. So then they say to make the narratives comparable, we need always some hero and some antagonist. And this is for example, you can say the hero is the technology for example. You can say there are so many people on this planet that we can change our behavior and consume or we will forced to change our consume. But the hero, nutrition of microalgae can save us from this conflict. What must be done that microalgae can save us? And then, oh well, let's think of, okay they may not produce so much carbon dioxide. Let's look up this and that's the way how already there's some interaction between this conflict stuff. So it's driven by this, yes. Any other questions? So it's okay, yeah. So I wonder if you could elaborate a bit on the feedback loop between the data analysis and the actually story. So question is, of course you want to like reform your story but of course, if you're looking for the truth there's this uncomfortable thing called reality which means your story may be off and you have to basically replace it or replace parts of it. So the feedback loops that are there, I think I got it, you mentioned this? I think I, yeah. Well, the truth in science is mainly true because it's peer reviewed. So this means there is some scientific community that peer reviews it and says okay, it fits to what we know so far or there are no arguments that we have to change what we know so far and so on. So there's no truth outside of this feedback. So all truth and all stable stuff is produced inside the systems. But this is some philosophical constructivist perspective and this is a system theory perspective so this is mainly well, the schools I was educated into but they interested me the most. But the thing is, if you look at narratives then it's also the history that you grow up with. So it's not just, then it's not just true because it's not falsified, it's also true because you can say okay, they did this experiment in the past and they tried this and somehow it's the evolution of the knowledge that lead us to this place so it's not relativism. And then it's not arbitrary and then this gives you some more stable ground. So this is maybe part of this feedback loop and the history of the feedback loop. So the constructivist also refer to this oral boros. This is the snake who always eats the tail of a snake like over there at the game. Yeah. Yeah, thank you for your question. Thank you for your talk. I would ask for one more plus so we can finish on time. Thank you very much.