Tekstet av Nicolai Winther Nå er det en av de fleste som har kastet seg på denne vegen. Thank you. Everyone, to the last session for the first day. My name is Michael and I co-chaired the research track together with Philippa Ski and Catherine Isbister. We received more than 30 submissions and out of those 30 submissions we selected six high-quality papers. And we'll start with the first contribution. It's called XR meets the performing arts. Sorry. Mobile phones. XR meets the performing arts led by laboratories at intersection of theater and virtual production. And this was offered by Ilya Mirsky, Volker Hetzle, Ludger Engels and Simon Spielmann. So please welcome on stage. Can you hear me? Yeah, obviously. So, yeah, hello to this afternoon session. So let's quickly start because I heard we just have 10 minutes time. So my name is Simon. I'm here with Ilya. I'm going to present you what we are doing at Film Academy Baden-Württemberg in cooperation with the Academy of Performing Arts, both in Ludwigsburg. I myself and Volker as well, we are from the Karl-Lemme-Institut, which is part of Film Academy. Film Academy, for those who don't know it, is a classical film school close to Stuttgart in Ludwigsburg in Germany. So there you can study classical courses like writing, acting, not but directing, camera, as well as visual effects and computer animation where we are closely connected to. We from film from Kallebner Institute we have an extensive background in virtual production and we're running since 2013 already EU-funded projects in the scope of virtual production. Basically, you can see it here on this picture on the left top, we already started in 2007 developing our own software around the scope of virtual production, virtual reality and XR technology. We are also running and supporting workshops like the workshop we had with the Academy of Performing Arts. Hello, good evening or good afternoon everyone. I'm Ilya Mirsky, I am lecturer at the Academy for Performing Arts. I lecture on theater and digitality. The Academy for Performing Arts is on the same campus as the Film Academy, where the Karl Lambert Institute is placed. And the Academy for Performing Arts, in German it's Akademie für Darstellende Kunst, was founded in 2007. And it has programs in acting, directing directing and a master's in dramaturgy and also a very strong focus on practice-based and interdisciplinary artistic education. There is a short trailer which you can also see on our website and I'm teaching there for five years now, Theatre and Digitality. And the paper we are presenting now is a collaboration between both institutes on this campus in Ludwigsburg. So, basically, at our Academy for Performing Arts, there is in the curriculum like an obligatory theoretical Monday Theorie Montag it's called and it's about introducing all the students at the Academy for Performing Arts to theoretical frameworks which they need for their education acting, directing and also dramaturgy and in the program on digitality or technology and theater, I am trying to teach them about the technical continuation of looking at theater as a history of technologies because basically the very first theaters or theaters were the very first theaters or theaters were the very first institutions who got electricity to run like the lights back in the end of the 19th century so you can see here it's like from a publication from the theater scientist ulf otto a very first sketch about a map in berlin and you you can see like the very first electric cables were going from the power plant straight to the theater and to the to the castle so back then like having new technology and theater was something very common and what I do is I focus there on looking at technology and media technology especially as something which highlights performing arts instead of like giving away the focus to the technology because in performing arts is all about the human actors and the human body and not so much about the technology and that's why and how we got together for the workshop for LEDVol Lab. Now some words to the technology or maybe where we're coming from. So first of all, what is XR? XR is a catchphrase for everything which is basically under the roof of AR, augmented reality, mixed reality and virtual reality. under the roof of AR, augmented reality, mixed reality and virtual reality. And we see also virtual production as in this world of XR technology. Who of you know about virtual production or exactly knows what it is? Raise your hands. Oh yeah, quite a lot. So basically here a small quote from the Visual Effects Society. So virtual production is a collaborative and So basically here a small quote from the visual effects society. So virtual production is a collaborative and interactive digital filmmaking process which begins with virtual design and digital assets, develops and continues in an interactive nonlinear process throughout the production. Interactive and nonlinear is the important thing here. So because in the last decades what happens to film production is that you often on a film set just see blue screens and green screens. And the idea, one of the many ideas behind virtual production is to bring back the filmmaking like we did it in the past. You can see it here, a good example of Metropolis where really almost the whole film has been made on film set a whole film crew saw or sees what happens on the film set of course there is still in this was post-production there is a cutting day is sound etc but the intensive amount of creations, foregrounds, digital backgrounds, all this stuff does not happen in virtual production in the post-production, but mostly directly on a film set. So the teaching format LED wall lab, which we developed throughout the last teaching format LED wall lab, which we developed throughout the last three years, has the core challenge how to work in the context of performing arts with a huge LED wall and it's asking students and those are all first-year students to use the LED wall as something which can extend the human body instead of like taking away the attention. So for that the students get like an introduction into the basic technologies and concepts of Unreal Engine and also about like a theoretical framework where this whole technological background is coming from and where it's actually aimed for, as Simon just said, more for film and media production purposes. So we think, and especially I'm convinced that in the future, LED walls will play a major role in theater. So it's actually somehow maybe XR in that sense is the continuation of the electrification of theater. And so we also have the opportunity to showcase small performances the students produce within a very limited time frame. So they just have two hours of working with the LED wall. So most of the stuff has to happen in advance conceptually and also really focusing on how to kind of incorporate the LED wall as part of their artistic performance. So there are always groups of five to six people, some actors, a director and a dramaturg who develop together with students from the film academy like those artistic concepts. You will see a trailer in a moment. Sometimes those students also decide to use depth sensors also to somehow make the LED wall more interactive like Kinect or Intel Sense. Here on the right down corner you also see like a very special moment because we have also students with disabilities in our programs so there we also try to make another use of the LED wall as something which amplifies certain movements or somehow, like in this case, it's an acting student and she cannot see, so we somehow play with the brightness of the LED wall to somehow create a different tension in this artistic performance. And in the end, after those students develop for several weeks their performances, And in the end, after those students develop for several weeks their performances, they are all highlighted and showcased with a discussion afterwards in the studio, but with a very restricted amount of time. So, yeah, absolutely. So like I said, virtual production, all these technologies around it, is way more than just technology for movie making. It's more a sandbox of tools you can use in a lot of purposes. And we also have the luck to announce that we will have a permanent virtual production installation, including a LED World Film Academy, which will be then also available of course for the students of the ADK. And yeah, let's see how we can further develop the collaboration here. Yes, so I think there's definitely a lot to explore because theater is like one of most of the ancient art forms has like a lot to learn from new technologies and especially I think the most important thing is always to create artistic works which really highlight, as I said, the performers and how this can look like you will see in this short trailer about this year's presentation. Die Aprikosenbäume gibt es, die Aprikosenbäume gibt es, die Fahne gibt es und Brombeeren, Brombeeren und Bromen gibt es und den Wasserstoff, den Wasserstoff. Die Zikaden gibt es. Wegwarte, Kronen und Zitronenbäume, Zitronenbäume, Zitronenbäume. Die Zikaden gibt es, die Zikaden, Zeder, Zitresse, Zerebellum. Die Tauben gibt es. Die Träumer, die Pucken, die Töter gibt es. Die Träumer, die Puppen, die Töter gibt es. Die Töter gibt es. Die Tauben, die Tauben, Dunst, D den Tod und die Gedächte gibt es. Die Gedächte, we end our presentation. Thanks for your attention and looking forward for your feedback and your questions. Yeah, thanks for this great talk. Are there any questions from the audience? Anyone? Okay, I've got a question. So in your paper you mentioned that each student group had a limit of about two hours within the wall. And you say that this forces creativity. But if the students would have more time, how would you deem that the effects will be? So I think it really depends on the students, of course. I think this limited amount of time has something pressuring, which is really like a value of this pragmatism, which is incorporated into theater makers. And I think, I mean, of course, having more time would mean we could also play with virtual cameras and really use more elaborated VP concepts for kind of extending the possibilities of this LED wall. So not just using this interactive part or like a big screen and a bright screen, but also more interactively. And I think this would mean we would need a lot of more time. So I think it's either or, and that's kind of a question of scalability. So I think either's either or, and that's kind of a question of scalability. So I think either it's kind of one week playing with the LED wall and the whole virtual production set up and really making a whole like play with it, or like staying with this little amount of time, which is somehow narrowing down the possibilities by forcing and fostering creativity in that way. It really helps in stopping thinking about too much about technology. You're just there and then please do it now and don't think about the surrounding. Of course then you need help from people which are running the system, which are there but this also gives you a bigger freedom of just doing things instead of thinking about what could it be and what happened if we and so on. You see it quite often in theater. There are young directors who are really talented and then they get the chance to stage something on the big stage and use all the technology which is there. Those plays sometimes become boring because they want to use all the technology because they have all those resources. And I think there is something in this moment of limiting those resources but still having like a strong concept which can be realized in a very short amount of time and speaking about theater nowadays it's also kind of something which is very realistic going out of the Academy having like a very less and less well-functioned subvention like cultural scene. So working in theater is always kind of working with limited resources in that way. Thank you. Any? Yeah, we have one question. Thanks. Really interesting, and I'm just, maybe this picks up on the last thing you said about giving them two hours where they can experiment, but they don't have too much time to experiment. I was wondering if you talked to the students about the potential danger of the technology overshadowing the content, and if that's an issue that you discuss, and of course, how you counter that. if that's an issue that you discuss and of course how you counter that. Yes, thanks for your question. So in the preparation to those workshops and to those presentations we speak a lot about technology and there is always and it's getting like year by year quite more difficult to talk to the students because students who go to do theatre are quite often very sceptical about technology. That's the reason why they go into this very analogue art form. So it's very important to somehow have this onboarding process convincing them that it's really about trying to keep up to a certain kind of tradition and history of theatre as a technical art form where it's always about the human. So it's kind of interesting in regards to that when theatres became electrified, like there was this big problem that the humans disappeared so they had kind of the problem to somehow become more present because the light was so bright in the back that they needed some spotlights. So spotlights were just introduced after theatres were electrified to highlight the human body. So there are certain lines of tradition, that's what I want to say, and I think it's really important to make everyone aware if you go into like this professional career of becoming like a theatre actor or a theatre director it's really important to somehow be aware of that theatre is not on purpose so analog how it is today but it's because there are so many resources from the past 56 years which were not used in the way how they could have been used so now it's a new kind of entry point to use virtual production because also through virtual production film production becomes way more theatrical everything is somehow centered in the film studio with like a LED wall and by that also film production is again way more theatrical because it's all in the set, right? So there are very interesting movements and shifts happening right now. Okay, any other questions from your side? Yes? Thank you, very interesting. At what point of production do you use the screen? Is it more sort of an idea exploring at the beginning of the creative process or more towards the end to test if things really work out where they have those two hours? So from our side it's a classical film production. So we really use it in the scope of a film production as virtual backdrops, as surroundings and even for virtual foregrounds and all this stuff. So it's a classical movie production or commercial production in the scope of Film Academy's teaching. And we are using it like as a very special medium and as a stage design somehow as a part of a very important part of the scenography which somehow can as in a regular theater play like become a certain part of the way how performances or stories are told and presented on stage. Does it change a lot after you've projected it? Or is it sort of what is there is there and you have to have produced it before? This is what we like to achieve. So what is there is there, so and everything is called in camera and then everything is done in best case. Of course, there are pros and there are pitfalls, but this is our aim. Thank you. Okay, yeah, I think... One more question. One last one. Thank you for your talk. Do I understand you correctly that by now, using LED walls in theater and also in film that the stage set in a theatre and the LED film set are getting closer and it's basically the same way of production. Is that correct or what did I misunderstand that? So that's a very good question. I think I'm not, I don't think that's getting closer but i think it's kind of similar yeah i think uh i think the process is becoming more theatrical so film production is becoming more theatrical through virtual production and that's kind of interesting because it's not separated into pre-process like pre-ning, like shooting and then post-processing, so it's way more condensed and I think this is something very theatrical because in the theatre you have all the different departments working in one space with all the technicians, all the artistic department trying to somehow make this performance happening and I think this is something very special and I think in that regard film production can learn a lot from theater. Okay, yeah, so Simon, Ilda, once again thank you for your great talk. We move on to the second talk with the title, The Extractive Image World, Artistic Counterpractices in Gaussian Splitting, Nerves and Volumetric Machine Vision by Tom Milnes. Hi, everyone. Can you hear me? Okay. So, yeah, thanks for all coming. My name is Tom Milnes, so I'm fromf o'r Celfyddydau Cysylltiedig ar FFAMWYTH. Rwy'n edrych ar y byd ffyrdd ddysgu. Felly dyma brosiect ymchwil. Rwyf wedi edrych ar ffyrdd newydd o ddysgu spatiol, fel sgwrs yn sblatio nerffau a ddysgu ffoliometrig. Felly, fe wnes i gyhoeddi, neu fe gyhoeddi, papur yn ddiweddar, ganddo'n enw Splatstraction, a oedd yn edrych ar y esthetig a'r ecologi cymdeithasol, a'r theori, y archeoleg cymdeithasol, o ffyrdd newydd o ddiffyg ymddygiadau ymlaen, a oedd Sblatstraction yn ennill y broses ymarferol rhw'r ddau rhan, sy'n ymwneud â'r ffordd, yn y ffordd ddiwylliannol, o ran ystradio data, ond hefyd yn y ffordd amgylcheddol. Byddaf yn edrych ar hynny heddiw, ond mae'r peth yw edrych ar artistiaid sydd wedi datblygu rhagweithiau cydweithredol gan ddefnyddio'r fforddau ystradio'r ddau rhan, neu ar leiaf dangos ffyrdd a ffyrdd o weithysgu ymddygiad ymddygiad, neu ar lai dangos ffyrdd a ffyrdd o weithio'n gwych a allai eu cynnig er mwyn cyd-ddychmygu ymarferau ymddygiad. Yn y papur, roeddwn i'n edrych ar John Gerard's Spirits, Ars Electronica Future Labs, Influence Ground Truth, SkanLab Projects, Framerate Desert Pulse, pôl, a phrensig, ymchwil, a'r gwaith o......Morosiannau Aliyari, gwaith, isis, hefyd. Felly, mae'r hyfforddiant hyn yn pwysio'r ffantasi o ddiogelu captur realiti. the fantasy of a neutrality of reality capture. Often all these 3D scanning technologies are called reality capture. By making visible the sort of petrochemical residues, the scanner agency, environmental duration, the evidentiary uncertainty in the cultural extraction, and infrastructural costs of these kinds of images. So the paper argues that probabilistic spatial images are extractive image worlds, a'r costau ar gyfer y ffyrdd o'r ffyrdd hyn. Mae'r papur yn argyfwng bod ffyrdd ysbasiol yn fforddau ffyrdd ffyrdd ysbasiol, ffyrdd sy'n ysbasi amgylcheddau neu, o ran ecologaethau, ffyrdd, materion, canolfannau data, dŵr, gwaith, a chwastraff. Felly, yn amlwg, dros y 30 mlynedd diwethaf, mae yna ddysgu fawr o ran ffynu ysbasiol. Ar hyn o bryd, bydd llawer ohonoch chi'n ymwybodol â sganio 3D, ac wrth gwrs, ffotogramedriaeth. Felly, mae'r argwmant yn y papur yn ymwneud â'r ffynu ysbasiol ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r ffynu ymwneud â'r fwyaf yn ymdrech yn llai fel llyfr nag ydy'r byd ddiddordeb. Mae'n dod yn ffordd o'r ffordd ysbrydoli, edifio a chyflawni ysbrydol. Mae ddiddordeb fel gwrsion, splatiau, nerffau, adnoddau 3D, a systemau'r gyfnodol am ddod â'r amgylchiadau hyn i'w gyflawni i mewn i ddynion ymdreol sy'n gallu cael eu gwylio, gweithio, edrych, ei ddysgu ac ei ddatblygu ar draws llwythoedd. Felly, yn dylunio hanes diwylliannol neu XR, mae hyn wedi cael ei ddynu fel captur realiti, yn aml, neu mae'n rhan o'r byd captur realiti hon. Felly, mae'r gwasanaeth yn amlwg o ran y cyfathrebu cyffredinol rhwng y byd a'r ddynion, ond mae'r llwyth o'r byd dynion yn dibynnu ar ystraddoiaeth o'r byd ffrindiau yn dibynnu ar ychwanegwyr sydd yn anodd i'w ddefnyddio ar y pwynt defnyddio. Mae ffotogramedigaeth, LIDAR a chynllunio llaw sy'n seiliedig ar y ffotograff cyffredinol gan gyd-dodd ymlaen mewn mesurau, clywchau pwynt, yn gwneud dyfodol o ddifrif. Dyfodol yn rhywbeth a oedd wedi'i benderfynu, ei gyflawni, ei gyflawni ac ei gynhyrchu. Ac mae ddifrifion fel Gawson Splats a NERFs yn gynhyrchu'r cyflwyniad hwn. Felly mae fy ymchwil wedi bod yn enwio'r fath o gynyrchu o'r cyfnod o ddifrifion cyfnodol arno. Felly mae ddifrifion gwyddonol yn fformau ddifrifol sy'n gweithredu sy'n gweithredu fel cyfnod o ddifdwyliant spatiol trwy estimad cymhwysol, yn hytrach na trwy ysbais optig unigol neu safon geometrig stafell unigol, yn ymddygiadu'r ddiddordeb cymhwysol ar brosesau ysbais. Felly, yn ddiogel, rwyf wedi bod yn edrych drwodd, dyma fframing archeoleiddio cymunedol, sy'n gallu helpu i ddeall sut yw fframingau gwyddonol spatiol probabilistig nid yn unig yn ffwrdd sydd wedi'i gynnal o ddechrau technolegau gwyddonol hwy, ond yn ymgyrchu a chynyddu ymddygiadau hanesyddol ar gyfer cymorth dimensiynol. Mae'n bwysig i'w ddiddordeb yw effeithiau dyfodol sy'n trefnu dimensiynolrwydd o'r ffotograffig i'r ymarferau stereoscopig o edrych ar y ddewis transplen i ddewis cyfrifiadau cyfathrebu cyfathrebu hefyd, yn dangos nad yw dyfodol yn unigol. Mae'n gysylltiedig â ffyrdd o sefydlu cyrff, lles a gwybodaeth. Mae ei gyfrifiad yn bwysig yn enwedig oherwydd mae'n cysylltu â dyfodfyrdd o sefydlu cyrff, ysgolion a chwybodaeth. Mae ei gyfrif yn arbennig oherwydd ei fod yn cysylltu â'r dystyr sy'n seiliedig ar y camera i ddylunio cynlluniau, i gynhyrchu, cyfrif, cynhwys, a chyflawni'r bydau gweledol hyn. Felly mae ime sbaethol probabilol yn cynnwys y hanes hwn. Nid yw'n cyflawni ffyrdd yn unig. Mae'r ddynion yn gwneud gwneud gwneud gwneud. Mae'r adroddiad yma yn fizial yn rhoi llwybr theoretigol yn bwysig i'r papur. Felly, nid yw imeion gweithredu yn unig. Nid yw'r imeion yn gweithredu i'r cyfrifion dynol. Maen nhw'n cael eu defnyddio fel rhan o broses o ran cyfrifiad neu systemau eraill. Maen nhw'n imeion sy'n gweithredu, sy'n arwain, dategu, clasifio, targedu neu'n hyfforddi. Felly pan fydd gwasganiadau neu nerffau wedi'u cyflwyno i Gorzyn mae'r syniad hwn yn newid ymwneud o'r ffermio diweddol i'r gweithgareddau sy'n y tu ôl ac sy'n cynhyrchu ac yn ei ddysgu. Mae ffyrdd yma'n cael eu defnyddio i ddarganfod y ffermio diweddol y gwelwch. Felly, mae sên ffermio neu ffermio radiaeth nid sydd gennych chi. Felly, ystod y ffordd ysbwll neu'r ffordd radiaeth yw nid yn unig ysbwll ysbwll, ond set o hyderau o fewn camera, cyd-dynllunio, dysgu cyd-dynllunio, data sets, gweithgaredd graffig a gwasanaethau platfform. Felly mae'r byd ime yn gweithredol oherwydd ei fod wedi'i adeiladu i'w wneud neu'i wneud trwy. Ac yn amlwg, mae'r hyn yma wedi'i adeiladu ar ystod ystod hir nawr o d ddewiswyr cymdeithasol yn siarad am y canlyniadau ecologol y cyfrifion technoleg ac yn sicr cymhwyso. Felly, ymddiriedaeth geolegol Eusiporica, ymddiriedaeth ffinitaidd Sion Kubitz, ymddiriedaeth cymdeithasol Jennifer Gabrys, Lisa Parks, a'r hyn a allwch chi ei gychwyn, yn ogystal â gwaith Sharon Matynith Sharon Matyn ar gyfer sefydliadau cymdeithasol y cymdeithasol, mae'r cymdeithasau'n nodi nad yw'r cymdeithas yn gallu cael eu llwyddo o'u cefnogaethau a'u sefydliadau. Mae hyn yn ddiddorol i'w ddod i'r pwynt o ddiflwyr spesial probabilistig, gan fod y cymdeithasau sefydliadol yn cymryd rhagor i brosesau dysgu maskin, hyfforddiant GPU, brosesu gweithredu ar gael, ac obsolesnwm ar gyfer hardwyr sy'n cynyddu. because they are increasingly tied to machine learning processes, GPU training, cloud-based processing, and accelerated hardware obsolescence. The aesthetics at the front end of a smooth depth and immersive scene are produced and obviously tied to rough material geographies. So as I mentioned, this is a dual extractive turn, geograffiaethau materol. Felly, fel y dywedais, mae'r troi'n ymwneud â'r ddwy-gynhyrchu. Yn ymwneud â'r ddwy-gynhyrchu, byddai ysbesiadau arddangos yn cael eu deall fel bydau ymgyrchu. Ac mae hynny'n ddwy-fynd, fel y dywedais, y cynyddu o amgylcheddau drwy'r cyfnod, drwy gyflawni cerfyniadau, cyrffau athrawiaethau i ffyniadau cymhwyso, felly dyna'r casgliad o ddata, ond hefyd o ran eu llwyth o ffyniadau o'r byd a'r amgylchedd. Yn ail, mae hyn wedi'i wneud trwy gweithredu ymchwil, camera, senswraethau, GPU, batreau lithium, sefydliadau cwylio data, yn ogystal â gwaith gwaith a chyfnodau e-bwyd. Gallai pob un ohonyn nhw fod yn ymwneud â'r penderfyniadau amgylcheddol y mae'r ffyniadau ffyniadau hyn yn arferol arnyn nhw. Felly mae'r ddiddordeb ddiddordeb y ffordd yw'r ffordd yma, y ffordd cynaliadol o ddynion cyffredin. Felly mae'r cyfraniadau'r papur yn ymwneud â ddim yn ymddangos bod ym, sefydliad gweithredol neu gweld yn newydd, ond mae'n ymwneud â sbesifio sut mae ffyrddiau ysbesiol probabilistig yn ymgyrchu'r traedegion hyn a'u symud ymlaen at gyfnod bywyd. Pan fyddwn yn siarad am y gweithgareddau cyd-drech, mae'n ddigon dderbynol i'w gwybod yw Jon Gerrard's Spirits. Mae'r gwaith gweithredol yn gynharach yn ystod ar-lein sy'n cael ei gyflwyno yng Nghymru drwy 25-26 oed. Mae'n gwaith sy'n seiliedig ar broses ar-lein sy'n cynnwys gwaith gwasanaethol, gwasanaethau gwasanaethol o ddoddau plasti, sgwrs, sy'n cael eu canfod ar beichiau o amgylch y byd, a sgannu ffotograff, felly mae'r ddoddau o''r rhain y gall y gwylio ddewis ac ymdrech â nhw, caelwyd mewn sblatiau gorsiwn o ddynion uchel. Mae gwylwyr yn cydnabod y ddynion hyn fel symuladau sculpturol, ac pan fyddwch yn clicio arnyn nhw mae'r ffwrdd o'r sgiliau petro-chemig yn ymgymryd â nhw. Y mwy y gwelwch chi â nhw, y mwy y byddant yn cael eu llwythu gyda'r gwaith olyg. Mae'r gwaith yn ymwneud â'r sblat i hanes petrochemig sy'n dod yn anodd. Mae plastig fel ffosil, y llwyth o'i llwyth o'i llwyth, a'r cyflawniad o'r môr fel cyflawniad planedol a'r symulad o'r brdysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ddiddordeb i'r ddysgu, mae'n ychydig o'r ffricsiwn, ond yn gweithrediad dechnolegol o astudiaeth materol sy'n ymwneud â chyflawni. Mae gweithrediad petrolewyr, cynnyrch cymdeithasol, drwy'r ddrin maraidd, ymwneud â'r cymig, a'r ffwrdd. Mae'r gwaith yn digwydd yn ddiwylliannol fel ddiwedd arferol o'r plastig a'r gwasanaeth cymdeithasol. Mae'r arfer GERWA yn bwysig oherwydd ei fod yn ysbrydoli gwahanol gwahanol ffyniad. Mae ffyniadau sbaethol ymddygiadol yn gallu gwneud ddiffyg amgylcheddol yn ddiffygol ond dim ond drwy ymddygiad yn y ecologi cyfrifiadol cyfrifiadol sy'n gweithio hefyd i ddychrynwad amgylcheddol. Nid yw'r ysbrydol yn llwyddo'r cyfranogwch. Mae'r byd imeiddio yn dweud bod yna ddwylyddoedd petrochemigol a threfn cyfathrebu. Gall llawer ohonoch chi fod yn ymwybodol â phrosiectau architechtiaid ffrensig. Maen nhw'n gweithio ar gyfer rhoi ymwybyddiaeth a'r gwir, projects, obviously they work providing voice and truth or they aim to provide testimony and show reconstructions of voices against oppressive or totalising regimes that may be obscuring evidence as well. So they do this in a very visually engaging way so I think their spatial investigations use yn defnyddio 3D ffotogramedig, sylfaen ffotograff, sylfaen fideo, a chynhyrchu architechtol, yn ogystal â chynhyrchu ardal agored o bobl ar y llaw trwy ffotograffu a chyfrifoldebau cyhoeddus o ddewiswyr cyhoeddus, a diogelu amgylcheddol. Felly, mae eu gwaith yn bwysig oherwydd mae'n trafod adeiladu ymddygiadol yn ddim fel realisme ymddygiadol, ond fel broses ymddangos sydd wedi'i ddynnu, ac mae'r ffyrdd, ymddygiadau a'r llwybrau yn rhaid iddynt fod yn ymddangos. Mae'r ffyrdd yma'n ddibynnol i ddeall gweledigaeth ymddygiadau ymddygiadol, wrth ddangos nad yw adeiladu 3D yn unigol, ond yn ymddangos, ym 3D reconstrucio'n unigol. Mae'n gyfadwyol, yn gyfadwyol, yn sefydlogaidd......a'n gysylltiedig o fewn y llwyth o ddewis. Mae eilweithredwyr A.L. Wiseman yn ysgrifennu ar weithredwyr. Mae llwyth o ddewis yn arbennig yn arbennig......herwydd ydym yn ymddangos i ddewis ysbasiadau spasiol......a'u gwneud i rhanau o'r ddau ffyrdd yn ymgymryd â'r ddau ffyrdd a'u gwneud i fod yn ymwneud â'r ddau ffyrdd. Mae gwirionedd poliwng amgylcheddol yn digwydd yn aml ar sgailau sy'n cynyddu'r gwylio, a'r amgylcheddau'n cael eu llwyddo, ac mae modd i'r modelau amgylcheddol gwneud y ffyrdd fel yna fod yn ddigon ddiffygol, ond dim ond os yw'n gwerthfawr a'r ffordd y mae'n rhan o'r dyfyniad, fel gyda gyd-architechtiaeth ffrensig. Mae gyd-architechtiaeth ffrensig yn cyd-arhau ar ddyfyniadau rhanbarth ac yn dangos sut y mae ffregion yn cefnogi neu'n herio cyfrif o ddydd. Mae hyn yn bwysig i gwrsyn, sbatyn a nyrs, oherwydd gall eu tecniadau gynhyrchu'r awdurdod o ddyddwch. Gall byd y ddyddwch yn ysgda yn ymddangos yn fwy cyffredin na'r ffordd sy'n ei gwerthu. Mae architechtiaeth fferensig yn cynnig epistemoleg yn y gwahaniaeth. Bydd y ddylunio'n ddisglos ei ffyrdd, ffyrdd, ymddygiadau a'r llwybrau. Nid yw'r problem yw bod y y cyflwyniadau'n cael eu cael, ond yw'r caelwch yn gallu cael ei wneud i ddod yn gyfan drwy dechnolegau a fydd yn gynhyrchu a chael eu gosod. Mae morosiawn a'r alwadau a'r isis yn gweithio ar ddyfodol 3D, gan sganio a chyflwyno i ofyn ar ar y gwleidyddiaeth o ddiogelu, gweithredu a chwltur. Rwy'n credu bod hyn yn bwysig yn benodol yma, oherwydd nid yw'n ymwneud â'r degradadau amgylcheddol rydw i wedi siarad amdano trwy'r prosiectau eraill hefyd. Mae problem yma o ran y gweithredu cwltur, a phwy sy'n cael ei sganio a phwy. Mae ei bracteisio'n bwysig i ddeall, oherwydd mae'n cael ei wneud gan ddysgu ystraddoedd, ond nid ar gyfer ymddygiadau amgylcheddol neu gyfnodol. Mae'n ddysgu am ystraddoedd epistemaidd a chwlturol, sy'n ymwneud â phwy sydd ganddo'r hawdd i sganio, storio, cyrru, gwaith, neu ffynnu'u data ystraddoedd, o ddynion, llefydd a hanes. Yn ei ysgrifennu ac yn ei gweithredu ar ddigidolololwg, mae hi'n argyfwngi bod technolegau canfod a chyflawni yn gallu cynhyrchu cysylltiadau cysylltiedig, neu ailgynhyrchu cysylltiadau cysylltiedig. Felly pan fydd cyflogwyr yn gynhyrchu, gynaliadn cyrffio'r data o gwerthfawrau diwylliannol heb ymgysylltu â'r cymunedau neu'r hanesau o ran y gwerthfawrion. Mae ymarfer Alliari'n bwysig oherwydd ei fod yn cymhlygu'r rheterig o gaelwch digidol. Gall ysgannu ymddangos i resdod o hyd i bwyllgor o ddystrychu, yn sicr yn y achos yma lle mae llawer o'r bwyllgor hyn wedi'u dioddef gan ISIS. Ond mae'r problem hwn yn dod yn fwy fwyaf ar gyfer ffynciau probabilistig oherwydd bod bywydau'n ysgannu yn gweithredu yn gyflymach yn fwyaf am gynaliadau pan maen nhw'n cael eu gwylio, eu optimio, eu cymysgu, eu cymryd a'u defnyddio neu'u defnyddio ar gyfer hyfforddiant, fel mewn setiau data hefyd. Felly mae'r gwaith yma'n ymwneud â'r farn amgylcheddol y tu hwnt i'r carbon a'r gwaith gwaith, mae'r gwaith ymwneud â'r gwaith ymwneud â'r carbon a'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaith ymwneud â'r gwaith, mae'r gwaysig pwy sydd ganddo'r hawdd i'w ddod o hyd. Yn ystod y studiau, mae ystod ymddygiad cyffredin. Mae pob un yn gwneud hynny mewn ffyrdd gwahanol. Mae'r gwaith hyn yn helpu i ddiffynu'r byd ffyrdd ymarferol. Nid yw'n unig y ffyrdd ymarferol, ond ffyrdd ffordd ymarferol sy'n cael ei gynhyrchu. Mae ffyrdd ymarferol yn ymddangos ar ymardd o ddiffyg yn fform ddiffyg gan ymdrechion ymdrechol. Mae ffyrdd spatiol yn arwain ar ymdrechion data gweledol o amgylcheddau symuladwy. Mae ymdrechion y data yn gweithredu i ffyrddau a'r modelau.....a'r optimaliad o ffyrdd trwy gyfnod o ddiffyg yn yr aelodau. Ac mae'r cyflawniad o asetau trwy gweithdai a sefydliadau gweithdai. Mae'r bryder amgylcheddol yn cynyddu, yn ogystal â'r bryder gwleidyddol, yn arwain materion o ddilyddio, cyflawni, cyfweliad, captio'r dynol, ac wrth gwrs, y cyhoedd o symudiad symlwyr. Nid yw hyn yn golygu y bydd artistiaid yn ddewis pethau fel gwrsion, splats neu nerffs. Cwestiwn mwy gweithredol yw sut y gall y ddewisau gael eu'r ffyrdd o'r ffyrdd yna yn ymwneud â'r ffyrdd sydd yn cael eu llwyddo. Mae'r ffyrdd yna'n ymwneud â'r ffyrdd sydd yn cael eu llwyddo. Mae'r ffyrdd yna'n ymwneud â'r ffyrdd sydd yn cael eu llwyddo. Mae'r ffyrdd yna'n ymwneud â'r ffyrdd sydd yn cael eu llwyddo. Mae'r ffyrdd yna'n ymwneud â'r ffyrdd sydd yn cael eu llwyddo. Mae'r ffyrdd yna'n ymwneud â'r rhain yn anghywir ar gyflawni'r amgylchedd. Nid yw'r gestion hyn yn llwyddo'r byd ime, ond maen nhw'n ei chyflawni o'i gwneud yn ddiddorol. Ond, mae gwneud ymchwil yn ymddangos yn rhywbeth yn hytrach na chyflawni ymchwil yn ysgol yn wahanol na'r un fath o addrefnu gwasanaethau trwy ymarfer. Mae'r gwerth cyfrifol o waith fel hyn yn y lle, ac felly, nid yw'n gweithredu'r imogiaeth ysbasiol probabilistig neu'n canolbwyntio'r anodd amgylcheddol. Mae'r gwerth yma yn ymarferau datblygu sy'n draw ymwneud â'r newansiau o to the nuances of technological inference and ecological upheaval behind everyday imaging systems. Thank you. Okay, yeah, we have time for one or two questions. So anyone from the audience? Yes. Thank you very much for this very interesting perspective on image making I'm just wondering how can we measure this ecological impact you name a few stuff but how can we really measure it I think that's the whole problem is so obscured, it's obfusc's the whole problem. It is so obscured. It's obfuscated, the exact sort of costs of this kind of imagery. I really don't know how to answer that because I don't know how we'd actually measure it. I think whenever I've seen predictions or people's estimations, it does vary wildly and this is often because it is collated from various sources of data around neu estymadau pobl, mae'n gwahanol. Mae'n aml, oherwydd ei fod wedi'i gael o ran ffyrdd o ddata am weithio mewn gwahanol ffyrdd o ran costau extraciwn, gwasnau e-gwaith, neu'r amryw o ddiddordeb. Mae hyn yn newid yn ymlaen. Mae'r dechnolegau ffyrdd yma'n datblygu'n gyflym, ond maen nhw hefyd yn newid. Mae eu hyder ar amgylchiadau amryw iawn yn newid yn llawer. Felly mae'n anodd iawn i bobl ddewis y pethau hyn yn gyffredinol, ac yn rhoi ateb onest i chi am un sblat fyddai'n costio hyn am y cyfan, neu amrywiaeth o gwrp, neu amrywiaeth o ddŵr. Rwy'n credu y byddwch yn cael ymdrin yn anodd mewn gwirionedd. Felly byddwn i'n ddiddorol iawn i wybod os oes unrhyw un yma this much for the environment or a certain amount of carbon or a certain amount of water, I think you'd have a hard job actually measuring that. So I'd be really interested to know if there's anyone here who has answers to that. But I think it would have to bring together so much to really answer it. Any other questions from your side? Okay, then. Thanks, Tom. Thank you for your great a great talk and yeah, we'll move on to the next presentation. So our next talk will be Generative Artificial Intelligence Generative Artificial Intelligence for Non-Representational Arts, Laura-Based Style Learning on Flux, by Jan Tietz, Niklas Beuter, Abdullah Sassiz, Al-Hamandi. And, yeah, welcome on stage. Yeah, thank you very much. And fortunately, I took the right microphone, so you can hear me. I'm Niklas, so I'm presenting work from a student of mine, Jan Tietz, and a colleague, Alhamman Adi, about generative artificial intelligence for non-representational art. I also call it objectless art. So this is one example, what you can see here in the background. So this is what it's all about. It's about art where you can't see any object in it. So I also call it like I said objectless. And what we try to do here is we wanted to train an artificial neural network to mimic this kind of art which is coming from a real artist which is known for me. This is what it's all about. This term flux here, this is the artificial intelligence model what we used for this. In 2025, I was approached by the artist who came to me because he read about artificial intelligence. And this is one real example from his art. So he painted this and he tried to copy this with AI. So his idea was, it takes quite a long time to create such a picture because he creates it, changes it, starts completely again and so it took quite a while and the idea is to use AI to be faster in that process. So he asked one AI model in that time, please create an image with yellow stripes with darker ends on a background and similar prompts but all what he got was something like fries or sometimes bananas. So that were things the AI model was known or knew at that time, but he was not able to recreate art in his style. And then he saw me in an article and approached me and asked if I can do that. And I said, oh, this is not that easy because AI models are trained on data from the internet and your art is not very well represented there. So let's take some students and train a model for that. And I asked the students, so are you able to generate an AI model which is able to do that? And then at the end, are humans able to differentiate? Is it real art? So perhaps real is not the right word, human-made art. Or AI art. And at the end, what does the artist say? Is he satisfied or not? So what the students did, they took original art, which is in the top line, and then this flux baseline model, and they first tried prompting. But they were not successful at the end. So they took a learning method, Laura, I think we already heard it today, where I don't go deeply into how it works, but it is a very efficient way of training a neural network by presenting images and prompts describing these images. And so they tried here with the Flux plus LoRa to recreate these art styles from the top line. So the idea is not to exactly match what you see in the top, but to match the style, how it's created. And they were again not successful. They tried several ways in quite a long time, but at the end they recognized, so there needs to be something more. So they approached the artist again and asked him to describe his own images. And then they recognized that you don't only need a good model and some specific prompt, or we call it generic prompt, but you also need very specific vocabulary to describe this kind of artist style. And then again they took his vocabulary, so his description of these artworks, digitalized 127 of his canvas or painted canvas, and then they trained these models. Like I said, without this generic, sorry, with a generic prompting style and then with the style, with the vocabulary of the artist. And then at this point, to be honest, we were quite satisfied. So the style itself, like if you take the first image, is already quite close to the one from the artist. So you have this kind of art is created by moving color on the canvas with different kind of tools like, for example, scrapers. Or I don't know if it's exactly, I think it's called Windows Squitchy, something like this in German, it's Scheibenabzieher. So the colors moved on the canvas in different ways and that's how the image at the end should look like. And they also took one of the most prominent models in image generation, Nano Banana, and also tried the same there. So, and now we wanted to measure, is it good? Yeah, did they match the style of the artist? And we wanted to measure this in a quantitative, in a qualitative, and in a, yeah, expert meaning way. The first one, I'm sorry for that slide to be honest, so we don't go into all the numbers and details, I just want to emphasize reading the paper at the end, but the quantitative measurement did we met the style of the artist was measured with a lot of metrics at the end. And just to summarize this, indeed the image style of the creation of how the color is been shown on the image itself and how the features of a single image is compared to the ground truth, the style indeed was met quite well from the Flux plus LoRa artists where the artist grounded captions. There are some metrics where the other models are better, and this is more like, for example, in the base, if you compare, is the prompt fitting perfectly to the image? And this, I think I can explain, because our model is trying to fit the style itself, and for example, NanoBanana is really trying to follow the prompt but it's sometimes forgetting about the style so at the end we were really happy the numbers somehow showed that the yeah artist style is really met with our training and With this way of using a lower with this artist grounded captions Next we wanted to know the second question. Are humans able to differentiate these kinds of art? So, we made a public exhibition, and 70 out of these visitors answered a questionnaire, and we showed different kinds of images. So, art made by AI and art made by our artist. And all of them stated they will be definitely able to differentiate these images. But if you look at the numbers, how many of them were able to differentiate correctly, it's close to chance. So they were not really able to differentiate AI and human-made art at the end. And the cool thing at the exhibition was that the students created also a website where the visitors could in a live way just enter some prompts and then the website created a new art in the way or in the style of the artist. or in the style of the artist. And the third question has been, so is the artist satisfied at the end? So what does he say? And he was indeed also very satisfied with the kind of pictures which were created. So he said that like the brush strokes and the textures have been pretty close to his style, but what he mentioned was that there are still physical gaps. For example, if you let the AI create an image, it forgets about resistance on the canvas, for example, about the physics of the color, how it sometimes fades on the canvas and so on. So if you're an expert and really looking closely at the images, you still can see the differences, for example. And then there's a very difficult question at the end or a statement at the end where we discussed quite a lot with all the visitors. He says that the style or the images are quite stylistically convincing, but not authentic. But he sees them as a good starting point for his own images to just try out things and then start from that point on. We had a long discussion with all the visitors about that, because they say art needs some kind of historic moment behind it. Some artist who was creating it with a story and something in his mind, or her mind, sorry. And this is where we discussed quite a lot, because this can also be created by AI. And if they don't know the artist in person, then all the story can be created at the end. So here are some results from the AI generator and yet to be honest there is one between which is from the human-made artist style. So I don't know if you can detect it, but that's it about my talk. And so thank you very much for listening. Thanks for the talk. Yeah, we have already a first question from the audience. Thank you. this is great. I have a question regarding the vocabulary of the artist. Have you checked the vocabulary with another artist to see if the other artist would describe the images in the same way? Not with another artist, but we took books where you can learn how to describe art. And so we took these ones and matched. Fantastic. Thank you. Any other questions? Yes. Which one was it? I should have made a questionnaire here for all of them. I will tell you afterwards. Thanks for the presentation. did you receive any comment like distinguishing between the 2D the 2D bidimensional presentation of the images as you are doing in your presentation against for instance the 3D plastic aspect of paintings which are when you paint that's why you see the traces. Yeah, thank you for that good question. And it's exactly pointing out the limit of this one right now. So we printed also again the scanned canvas from the artist to have a fair comparison only in 2D. And only then it's by chance differentiating them. Because if you have the real art with all the 3D structures on it and the real color on it, of course, there's a big difference still. So our idea was to take a robotic arm next and let that one learn to push the color and everything and then have the decision again. But you're right, this is still a limit. Yeah. Any other questions? Okay, yeah. Shall I tell which one is from the real artist? It was in the top left, the orange one. I don't know if I can. I don't have the presentation here anymore. Sorry for that. One of our people. Just a second. It's really complicated. Very exciting. Should I just go there? The one on the top left, which you might have not expected. Okay, so thank you very much. Thanks again, what you very much. Thanks again for the great talk. Yeah, so let's get to the next contribution. The next contribution has the title Feel-Back-Loop, Effective Co-Adaption Between Humans and Embodied Intelligence Systems Through Reinforcement Learning. And this was done by Ghazal Hosseini, co-adaption between humans and embodied intelligence systems through reinforcement learning. And this was done by Ghazal Hosseini, Hanif Akhtaralab and Philipp Wintersberger. So hi everyone, good evening. Thank you for having me here. Today I want to present the paper that I co-authored with Hanif and Philip, which is called Field Back Loop, Affective Co-adoption between humans and embodied intelligence system to reinforcement learning so a short introduction of the artwork this research conducted based on an artwork that I created last year and exhibited in Ars Electronica 2025 in campus exhibition of the interface culture department. And Fieldback Loop is a robotic tail that develops its own movement language to interaction with human emotion. So this is a short video of this artwork. So the artwork has a camera that can take the facial expression of the human in front of it. And it generates some movement based on this interaction or facial expression. This is another video that is like we found it interesting, the interaction, then we started to write a paper about it. So this robotic tail detects facial expression and develops a policy based on this expression, which is classified the expression into two different groups, like positive emotion and negative emotion, that it works as a reward to the system that uses reinforcement learning for development of this policy. And then the interesting observation that we had in the Ars Electronica, it was that people started to engage with physically through their body. They move with it like they dance with it. And it was interesting. It was the start of this research. So if I want to talk a little bit about the interaction. So interaction is usually an interactive art. It's like one direction. Human gives an input to the system and system executes a predefined interaction or behavior. But if we use different kind of interaction that human can input to the system and trigger the system and also the system can influence the human in another way so it's like a two directional interaction that I want to talk about it. So I want to talk a little bit about this kind of interaction. So if we want to say that it's not just one-sided interaction and it's not a predefined interaction, there are some artists that they work on in the 90s. They started working on evolutionary algorithm in the media art. One of them is Carl Sims, that he developed evolved virtual creation, like a virtual creature. And this creature could develop its movement and its behavior through the interaction. So there was no pre-programmed or predefined movement or behavior of this creature, but it could develop itself in the virtual environment based on the interaction with the environment and human. So for example, in this artwork, he developed this model and this like creature that could swim after sometimes like a trial and error, and based on the interaction that it got from the environment, and it was able to swim after some time. And another artwork by Christos Amar and Lugan, who is my supervisor, it's evolved in 1994, and also they They had the same approach for creating like a the virtual a creature That can develop and evolve based on the interaction with the human for example in this Example it was a touch and feedback loop is also is like in the same category of like evolutionary media art or interactive art But it just the input is emotion and facial expression so if we want to consider emotion as a interaction or like a as an input to the system. Then we were curious about how people started bonding with technology, or how would be the relationship between human and technology based if the emotion is a medium. So then we searched about it, and we found some people feel some emotional attachment to the AI chatbots, for example. But in the chatbots, it's just a screen and a text space. You give an input as a text and you got an output as a text. But if we want to talk about more a meaningful interaction, if it's like, we can say that another channels of like another tool of communication that is like nonverbal communication like a body language or like the facial expression and also we can consider emotion as another way of communication and also movement as another way of communication. And then Rosalind Picard, that a computer or a system is truly intelligent, it should be able to not just process the emotion and respond to the emotion, it should be able to express the emotion and kind of have emotion. And the thing that we call emotional intelligence, have emotion. And the thing that we call it emotional intelligence, then we can, if the computer or the system has emotional intelligence, then we can say that it's really, truly intelligent. So here, emotion as a medium, if you want to consider emotion as a medium, then there are more critical questions that can arise. For example, if machine can perceive and respond in a different way, then we should look at it more critically in terms of the interaction design. And this question, this critical question also will arise that what if emotion becomes the medium through which a machine and a human co-construct their relationship? Then if we can say that, okay, emotion is a way of like communication and non-verbally, then there should be a way that we can say that the system can adopt to us because it's not just us that we can influence the system as an input, also system can influence us. Then another, in the feedback loop, another medium of interaction or establishing a relationship with the participant is the motion. If you saw the movie, I showed it last minute before, it was like the shape of the tail. And the tail is like, it's a metaphor of the communication non-verbally. like it's a metaphor of the communication non-verbally. Because in animal, they use the tail to express their emotions and their feelings. And also, they try to communicate non-verbally with us. And also, if you look at the more morphology of this feedback loop as a tail, it has like a vertebra, because we also, at the end of our spine, we have a tailbone, which is like in the evolutionary history, it's supposed that we also have, we had a kind of tail. And also there was an artist, a Japanese artist, Hachiya, in 1996 that he also developed a tail for a car, which is called a tank's tail, and he used this tail as a medium in the traffic, for example, to communicate with cars and traffic for example to communicate with cars and Nonverbally and like say thank you to the other cars that like were behind it in the traffic And This is the system architecture of the feedback loop as you can see We use the reinforcement learning in this artwork so the tail detects the emotion, the facial expression, actually, and considers it as an emotion. Then the algorithm classifies it to positive and negative emotions. For example, happiness is considered as a positive, which is considered as a reward for the system. And negative emotions, like sadness, are considered as a penalty for the system. And it's considered as a penalty. So it has a table. It's called Q-table or Q-learning algorithm that has a value. So each movement in the feedback loop has a value. So each movement in the feedback loop has a value. It starts randomly at the beginning and then it tries to evolve itself and develop a policy based on the feedback and reward and penalty that gets from the audience. So for example, if it start wagging, and if it gets like a more positive emotion, positive feedback, so it consider this emotion as like a positive emotion, and then updates the cue value and increase the cue value. If it gets a penalty and it decrease the cue value, and like it's less chance to use this movement in the next for example episodes or next encounters. And then we conducted a like like a user study and evaluated so we invited 30 students across four groups and with diverse backgrounds in the educational context and we allow them to interact with this artwork because like before and after the learning the basic knowledge about the reinforcement learning because we wanted to see that how they perceive it in like a three different dimension for example how they perceived adoption if they attribute any characters characteristic to this system and how also they interpret the movement of the system. And there are the findings of this, like this is the interesting finding that we found in the this user study and also from the observation in the Ars Electronica. So they said that for example it's unpredictable like a cat. So they kind of saw it as a cat sometimes. It wants something. It's curious. It's happy to see me. It's dancing. And somehow it's bracing and also sometimes they couldn't understand of course the, and they said that it's like ambiguous and random movement. So there are two different results. And I want to thank you, the Interface Culture Department that I study, and also ITU that I worked there. So and they I had this opportunity to conduct this research and go further with our work. Thank you so much. Thank you for the presentation. Any questions from the audience? Any questions from the audience? Okay, yeah so I have one, actually pretty funny one. Why a cat? So three or four groups saw a cat in the tail even though you never designed it that way. Did it surprise you? And would you now design the movement differently? More or less animal-like? It didn't surprise us because the first time that I accepted in the Ars Electronica, I also received a lot of feedback that they kind of like attribute some animal-like to the system. And because I designed it as a tail, so it was kind of like, you know, obvious for me that maybe they can see it as a kind of part of the animal or like an organic shape. Okay, yeah. I think I have a mic. I understand you designed it to optimize it for happiness as an emotion. Did you try other emotions if it would learn to make interactive people make angry or sad? Yes, actually it's a really interesting point, but in this artwork I just designed a goal as a like to make people more happy or go to the more positive emotions. But it's actually a really interesting point and yeah, I would like to continue it in this direction. Thank you. Other questions? No, I have an additional one. So what did the tail learn at the festival? After several days at Ars Electronica with hundreds of visitors, what did the tale's behavior look like in the end? So the tale was like at the beginning because it just used random movements and it's able to explore different movements to learn actually. At first it was sometimes really aggressive, like it moved really fast, but at the end, because this kind of aggressive or this kind of movement got more a penalty, so at the end of the Ars Electronica, it was more calm and gentle movement and smoother. the Ars Electronica, it was more calm and like, you know, gentle movement and smoother. Thank you. Okay, yeah, so thanks again for this interesting presentation. So we move on now to the next one. Big applause. Thank you. Okay, so the next talk has the title, Empowering Marginalized Narratives Through Pixel Art, a Case Study of the Video Game of My Life and the Medicine Association by Maria Paget, Nia Ehrlich, Ruben Tu, Lasse Löpfe. Okay, hi everyone, thank you for being here and thank you also the chair for organising all this and also the people who chose this proposal, we are very happy to be here. And I'm Maria Pagès, I work in the Centre of Digital Image and Multimedia in Catalunya. And there I work with Nia Erlich, Glaceleufe, and Ruben Tous, who comes from Computer and Architecture. But the four of us, we work in the Universitat Politícnica of Catalunya. So this project stems from the possibility of making video games with people who have not an easy access to video games. It's the title is the video game of my life and we just mix video game, digital art and as you see, social activism. We try to reconstruct the aesthetics of the 80s arcade machines and with some imageries that you will see that they are alternative, disident, and often invisibilized, because the participants of the project you will see, they are women, or they are women they are women or they feel like they are women and they usually, we use this word to designate female, transgender non-venerate and women of color, so these participants they are homeless, they are usually consuming drugs and you will see how they deal with the daily routines thanks to the video games that we will see. The project stems from Medzineras. That's a non-governmental association that's based in the Raval, which is a neighborhood in Barcelona, a very intense neighborhood with tourists, with people from generations that have been living in Catalonia, and also people from Pakistan, from Morocco, from everywhere. So they have this association that's for homeless and drug consumption women, as I said, and they have one thing that's very particular from them, which is they just don't try to rehabilitate these women. They don't try to stop them from consuming drugs because it's something that helps them to survive in the limits of the society. So they only try to give them a shelter, a safe space, where they can have a shower, where they can change clothes, and where they can have a safer injection site. So it's not a matter of just eliminating the drugs, but just to help these women to have harm reduction, because the limit between medication and drugs, it's really blurred in this context. So I get in contact with the association and I found the Symphonies of the City program. That's a program, that's a six edition, and it's a program to produce experimental shorts that's granted for the Center of Contemporary Culture in Barcelona, the CCCB. And they offer the possibility to make short films, experimental ones, six every year, and they offer the possibility to Donas Visuals Associates, which is an association of women that work in audiovisual domain, and I'm one of the associates, so I got granted to make this artistic piece of 10, 15 minutes. I'm not going to talk about this piece because it's in post-production. We will mix things about the gameplay you will see. We will mix the aesthetics of Pixlr and we will mix also the development of the workshops. But we will talk about the workshops and how these video games came out. Thank you. So I'm just going to take my minute to shed some light on some of the ethical issues that we had to confront with as we started our project. So first of all, our starting position was definitely a feminist stance. And since feminist video game studies, they aim to decentralize the male perspective and open up a stance for more diversity within the medium. So games are not just entertainment in this case. We're trying to create these lived experience, especially experience that are marginalized by mainstream media. Now, feminist game makers increasingly use interactive forms to explore trauma, memory, care, resistance, treating gameplay itself as a way of kind of engaging with these silenced histories. So the video game of my life takes up this approach and represents drug use and survival in a way that's meant to awaken audiences that are unfamiliar with this reality. Now our starting hypothesis in this case was that we had to protect the protagonist, which means that we had to maintain their anonymity. And that need really shaped all of our technical and kind of representational thoughts before we started. So we landed on this kind of double distancing stance, which was A, avatarization, the idea of using an avatar rather than seeing the protagonist, and B, of course, animation. And we approached animation in the sense of animation has this kind of dual capacity, this masking capacity, where on one hand it can hide because it hides the identities, but on the other hand it can also expose. It can expose the lived experiences, it can expose recognisable. It can expose recognizable locations. And it can definitely show a lot of the subjective aspect that otherwise can't be seen. Now, concretely, we used rotoscoping, which is the animating frame-by-frame of live footage through the tool Lester. My colleague Ruben will explain more about Lester. My colleague Ruben will explain more about Lester. And Lester allows participants to both animate and make a lot of the representational decisions. So it's also worth emphasizing that a lot of the underrepresentation is also a financial issue, which also keeps marginalized creators out of the field. And that context shaped everything, and also why we used Lester as a technical solution that meant providing maximum agency to these users, rather than requiring them to have funds and technological know-how. Now, the most important concept that we'd like to emphasize is that that sense of agency ended up really being key because these women seeing their stories rendered on their own terms they ultimately chose to be recognizable they consented to exposure and they move past the anonymity that we assumed they would prefer so as a result we're glad to share the images of the entire process with the women whose lives we depict in the game. Something a little bit more technical. The characters were designed with Lester. Lester is an animation editor originated from a research work carried out at our university at UPC. So the editor allows to generate an animation, as the ones you see here, from real videos, from live action footage, so it's a kind of rotoscoping animation software. The idea is to facilitate workflows such as the ones used by Jordan Mechner in the development of the classic game Prince of Persia i altres. Però, en comptes de anar a tots els frames un per un, havent de rotoscopar tots els frames, Lester et permet només de rotoscopar el primer frame i després de propagar les decisions artístiques a la resta dels frames automàticament per l'ús de diversos algoritmes. Un d'ells és un algoritme de segmentació de vídeo, però també algoritmes de simplificació geomètrica artistic decisions to the rest of the frames automatically through the use of several algorithms. One of them is a video segmentation algorithm, but also geometric simplification algorithms like Douglas Becker, for instance. Of course today you can achieve something similar with generative AI, as you can imagine, but it's typically more costly in terms of money and also impact in the environment. the process is not so deterministic. It's a local desktop application. It's free, it's open source. It does not require a lot of information. So, it's a very good process. Typically more costly in terms of money and also impacting the environment. Also with generative AI, the process is not so deterministic. artista e de controle sobre o resultado. Lester é algo tradicional, é unha aplicación local de desktop, é libre, é open source, non requere internet. Usamos-o sucesivamente en todos os jogos, os 4 jogos do proxecto, para o diseño de todos os caracteres, e tamén de todos os caracteres principais, e tamén de alguns dos NPCs nos jogos. Fizemos isto com a participación das mozas participando no proxecto. pieces in the games. We did it with the participation of the women participating in the project. They not only provided the performances for the live action footage, but also they sit with us in front of the editor, selecting the different colors and the finishing aspects of their characters. Maybe I'm going to talk about the workshop itself a bit, how we did it. And our goal, as Nia said, was to give agency to the participants. So we tried to give the freedom, the sense of freedom of the burden of having to implement the game. So we did that. So they just talked about their ideas they had and it was implemented by professionals. We did all the coding, it was not a sort of teach to how to make a game, as I said, it was giving the maximum artistic freedom to these women, which implied also very importantly the final word about what goes into the game and what not. So we could not decide on that as far as their decisions. but also very importantly the final word about what goes into the game and what not. So we could not decide on that as far as their decisions. So I want you to present the games resulting. Antonia started from Bubble Bobble and wanted to represent the place she lived during the first six months of her stay in Barcelona and all those ghosts, creepy creatures climbing up and she had to chase them out of the place where she was living with poison coming out of her hands. And the ghosts laughing at her. This is Maria's game. She is......represent using lemmings as a base game to show how drug addiction dictates her daily life. She feels herself as an avatar going through life and having to do all the stuff to get to the sugar daddy to get some money go down again buy the drugs be avoid being robbed so she can keep the money and the drugs and finally reach the missionaries Association to smoke her crack pipe in peace forced to do that every day and every day so it's one lemming after the other. Tania just wanted to destroy the MACBA. MACBA is the Center of Contemporary Art in Barcelona. And it's sort of a symbol of infrastructure done for tourists and rich people in general, but not the people living there. So her goal was to destroy it. And once it destroys, you destroy it, and once it destroys, what is left? Nothing. And there was a very peaceful nothing somehow. Then finally we have Kitty. Kitty was very inspired by the Catastrophe Anarchist movement during the Civil War, and shocked by the Catastrophe and Anarchist movement during the Civil War and shocked by the police violence she was experimenting on the street. So she wanted to be sort of a cat climbing around and jumping around in the balconies of the rubble and slicing policemen into pieces. That was the goal of the game. You see, they could choose what goes in and what not. The structure of the workshops was relatively simple. We just made them play some games. First session was just playing for fun. Second session, choosing from there what game to want to make, sketching that game out, then we had some recording sessions and tried to bring that into a first version of the game and let them see that. That was a crucial point because from the feedback we got, the games changed dramatically what was in the first version, what we have interpreted. So that feedback we implemented again, We did the final shootings for the animations you have seen before. They did some drawings. Some of the artwork is there. And finally, sort of a conclusion and feedback version to see what they thought about the process. What they thought was, I think, really positive. They really felt the agency to do the stuff and they were surprised because normally it's not something that happens in their life. Having an agency is something new for them and they really enjoyed the experience. So I really like the game because it's exactly as I had asked for it. It has all been very altruistic without anyistic, without any rules, without any obligations, not dictating how things should be, but rather do whatever you want and I love that. The video game was an opportunity to show our personality and from the mentor of Metsineros she told us that it was the first time in all the workshops that do that everybody is attending all the sessions. So I think from that point, it was successful. As conclusions, we were very proud of just proving that the routines of the medicine areas people were in the workshops. They decided to show. We didn't know if they would just make something superficial to avoid their real lives. But no, they put the routines, the things that happened to them into the video games. So I think that was really just putting real life into a ludic place. Then also, the process of design in the video game was iterative. We made, as Lasse said, we made seven sessions. We could just keep doing sessions and trying to polish the video games. And we decided that they had to focus on the narrative part, on the game design and not on coding as it's usual in this kind of workshops. Another thing is that they really feel very happy for the sovereignty they had, the possibility of deciding absolutely everything. That's something that we realize that doesn't happen in Helya in their real lives. And we're also very proud of having these games mechanics and this structural design from the decisions of all the participants so thank you very much for listening and we wait for the questions mess thank you for your great talk. Any questions from you? Yes. Hi. I actually have a comment and a question. The comment is that historically, arcades were a really dark place and not a very safe place in the 80s, at least. They were a place where there was mostly males and not really too many females or people from different groups. So it's some kind of poetic justice that you guys have built this. I love it. So thank you for that. And my question is, the gameplay design, the core gameplay design for those games was extremely difficult. And that was part of the engagement back then. And I'm curious if in the adaptation of trying to make a game for people who don't have the level of gameplay experience or gameplay skill, was it that difficult in order to create a compelling game, or how was that work process? Yeah, you're completely right. I mean, the first session, we've been playing the original games from the 80s, and it was madness. It was just, you were in the game over again, again, again, and very quickly. It was because they wanted you to put the coin inside the machine, so that's why the game mechanics were really hard. But in this case, we decided to transform this, not as we say video game, but they are mini games. And I have with me the teacher of game design from my center, so I will just let him explain the mini games mechanics. I think it was surprising to me the knowledge they had about the 80s game as well. So we started with the premise that they wouldn't know them, it would be difficult for them, but no. Some of them picked the game, Bubble Bobble, because she liked it in the session we were playing. Lemmings, we didn't even try it out, but she said she came, I wanted to be based on Lemmings, definitely. So she had that clear from her own knowledge as well, so shouldn't underestimate them. And one was Tanya, the destroying game, she didn't want any game as a base, it was all from scratch. Yeah, the last one is sort of a combination of a classical jump and run and Donkey Kong. She pictured it on the MACBA in the museum because there's sort of that diagonal stairs and she really saw, I see it there, I want to go up there. So it was the combination of the physical place helped a lot as well, no? Was the three things, the game starting point, their imagination and put it into a physical place is something I didn't say. Thank you. with the three things, the game's starting point, their imagination, and put it into a physical place. That's something I didn't say. Thank you. Other questions? Okay, so I have one. So the games that you created show crack smoking, cutting police people in two. Did you ever say no to an idea that was brought up? No. No, that's the whole thing. Did you ever say no to an idea that was brought up? No. No, that's the whole thing. The only thing we said no was technical limitations or to protect myself somehow. If you said, I want to do a multiplayer game, I said, sorry, no, I have a life. So at some point if it was sort of too hard to implement, or one side, I want 11 levels is no One level. Okay, so that was the only possible no, but no censorship at all I think that it was a reward for them just to see that everything that they said Maybe it was something stupid or maybe it was something that for them They they just love it, but when they saw the next session that it was a reality in the video game, it was like, oh, look at that, that thing, it's in the video game. So it was something that... It was nice not to say no to anything because it's something that they really enjoyed. I'll just add to that because I think it ties back to the ethical issues and I think that ties back to the Metsenderes goals. It's not about what you'd expect, right? It's not about cleaning them up or something like that. No, it's very much accepting their life situation and helping them within that and the game reflects that. Understanding more about their life situation and giving them the agency within that. Yeah. And it's very one to one. the animation of smoking a crack pipe. She was smoking a crack pipe. So it's it is true there. Great project. Any other questions from your sides? Okay. Yes, so once again thank you for your great presentation and yeah, let's get to the last speaker for today. Thank you. Okay, so yeah, the last talk has the title from screen to space, rethinking how interactive markers reshape physical digital interaction in theGames and this was written by Vinaya Tade, Nicole Dostalova and Simone Kegelstein. So welcome Vinaya on stage. Okay. That was weird. Cool. Hi. My name is Vinaya. And in this presentation, I'm going to be discussing about what happens when there's physical interaction that happens in mobile augmented reality games. Now, we've known that augmented reality is bridging digital content onto the physical world, right? But there's a real problem there, which is, although augmented reality is a hybrid version of reality, it's still very screen-centric when we talk about interaction. I can probably emphasize more about this problem through two interaction paradigms. There's the window of the world interaction paradigm where there is the world and the screen acts as the window through which perception and action is experienced by a player. And there is the window of the world as support, where the world acts as a place for interaction, where interaction is primarily the space where things can get manipulated, and there's movement and space as a context is used. At the same time, digital content is still superimposed into the space, making it more AR-like. Having said that, these two interaction paradigms help to create different kinds of interactions, depending on whether you want to focus on physical interactions or digital in AR games. Now, speaking about more interaction, I would like to maybe focus on embodied interaction because that's the whole basis of where physical reality comes into being. So embodied interaction is really going beyond the screen. So there is the space, there is the body, the objects and the artifacts around us that really help everything get embodied. And tangible interaction really helps create a very specific vocabulary for embodied interaction. So this is the space where tangible materials can get manipulated, they can be moved, the orientation can be shifted, and there is more space for agency, like you guys already spoke about, which really helps my presentation. So the space where our research comes in is, yes, a mobile augmented reality game can be a window on the world, but it doesn't necessarily have to focus and make screen-heavy interactions. We can move beyond and have physical space as a part of meaningful interactions. So markers can be one way of finding a solution to this exchange. Now as we know about image markers and there are QR codes that you scan and they give you augmented objects on your screen. And we already know about this. And there's recently a lot of work that is using intermarker relations and understanding how interaction works there. There's also a really interesting research done on stackable markers, where there are sort of transparent markers that you sort of stack on each other and they help you in their interaction. Now if you really see, even though we are talking about markers, these are all still 2D in nature. So what we understood in our systematic literature review is using 3D markers is still an unexplored space and it's still not being explored as a medium for using the physical world for interacting. Now this gives us an interesting proposition of using interactive markers as a configuration to help get the physical reality back into augmented reality. So our question was what kind of behavior patterns can emerge if players start engaging with interactive markers in a mobile augmented reality game. And second is when they are interacting with these markers, what are these interaction patterns that might emerge? So to understand this we created a research prototype, which is Roo. In Roo, it's really necessary to use physical construction to go ahead in digital progression. So a marker in this game in the first picture, you see you get a clue, and your riddle is to help understand which material to use and what to build. In your second picture, you see there's a silhouette. So you have to create a marker based on the silhouette and then you scan it and there's an augmented object that appears on it and that's how you progress. So it's an interaction loop that keeps helping the user educate on how to really move from the physical world to the digital world. Now we presented this at Expanded Play in 2024 and it was an exploratory study because we still don't know how users were going to interact with jumping in from the physical reality to the digital reality and back and so on. So we had about 12 participants out of the 70 participants who engaged with the entire game. 12 of them consented to the qualitative study. The sessions approximately lasted for about 20 minutes and the age range was between 19 to 45. We collected data from post-play questionnaires and there was a lot of discussion about it. So the first really obvious finding was construction of these interactive markers really redistributed the interaction from continuous screen device usage. So we're not saying that the markers were great and they got attention. It was important to know that the participants kept their phone aside while they were constructing objects and when they were done constructing, they picked up their phone and went back to play the game. So there was a continuous distribution of physical reality engagement to digital and so on. The second finding that we understood was recognition constrained the tasks. And when I say that, I mean the silhouettes that you saw, we needed to have them because we couldn't afford to give our players complete agency because we had recognition and computational issues because this space is still currently being studied. Therefore, we needed to give them constraints. But even in these constraints, what was interesting to see was the players still found a way to help themselves make their markers more customized. So in some of these images, you can see there are some markers that are being constructed horizontally, vertically. Some participants were actually completely rejected the table and started constructing right on the ground, which was very interesting to also see. Some of the participants also really appreciated using specific colors that you can see in the last image. And there were some participants who created extreme miniatures and very maximalistic versions of markers. And like I said, because this is a new form of interaction where you're jumping in and out of physical and digital reality, there was challenges that the participants faced. So this was more related to participants who were older compared to the ones who were younger, where they first tried to move their phones around in searching for the actual material, for example, bag. So they would go and try to scan the bag because they were unaware that they had to go in the physical space and engage with it. Now because their prior knowledge of mobile augmented reality games was everything is screen centric, they just happened to keep doing stuff on the screen and they were just unaware of being involved in the physical space. So five of them explicitly asked for help and the others managed to eventually figure out. So understanding these sort of behaviors that have emerged from physical digital loop, we have some implications that we derived when designing with interactive markers and mobile augmented reality games. So we want to encourage hands-on engagement, which is although markers interrupt continuous screen usage, they're not disengaging players. Second is to enable creative autonomy. So, recognition boundaries still persist because of computational limitations, but they still let players choose sizes, colors, orientation, and placement. And third is accommodating diverse interaction patterns. Now this is specifically with the age group. Depending on what age group you're working with, they might be more tech savvy or less. It's important to help them with digital cues and strategies that can help them shuffle from the interaction of physical and digital worlds. So to conclude my presentation, we don't argue that you need to be moving away from screen usage in mobile AR. What we're trying to ask or start discussions about is how do we change where meaningful interaction really happens? And that's my talk. Thank you. Thank you, Minaya, for this amazing talk. So, questions from your side. Okay, I have one. So, you mentioned that 10 out of 12 players first tried to solve everything on the screen and 5 needed help. Is it a problem you want to fix with better onboarding or should it be part of the experience basically? Is it something interesting for you? So when I said that 10 out of 12 were struggling, it wasn't because of the onboarding. They read it, they just did not anticipate engaging in the physical world. They didn't anticipate that they had to construct something in the physical world. They kept searching for puzzles around in the physical space, so more augmented objects, so they would move around, you know, trying to search for these augmented objects. And then when I helped them understand that you actually have to engage in the physical world, that was something probably novel, which they hadn't anticipated. And because they came with prior knowledge of AR has to be screen-centric, and we just move around with our screen. So the window of the world paradigm is where their focus was on. Hence, they just did not anticipate that they would have to leave their phone down, which was really surprising that they had to do that. Interesting. Yes. Any other questions? So I have another one. Everyone got the same materials in this control setting that you have. What happens when players use whatever they find at home? So does the scanning still work or will it feel the same? So is it part of this basic experience that you designed? I didn't have the answer to this till last year because people were just doing, you know, they had the materials in front of them. But I think last year when I was presenting the other game, there were these three really interesting participants. And they had to create, like, Lego markers. And somehow they saw a paper and started making Lego instead of, like, constructing them. And it actually recognized the pattern because it looked like the Lego blocks. They were extremely accurately doing it right. So I guess there are different ways of how you can play with it. Cool. Fascinating. Great. Yeah, so if there are no further questions, many thanks, Vinayaaya for this great talk and Yep, one big applause, please Yeah, so this concludes our last session for today thanks again for the fantastic talks the speaker gave and Yeah, so perhaps is a good idea to hand over to jürgen there he is thank you so uh first day of the expanded conference we will continue tomorrow 10 o'clock you probably know that there is an opening tonight and another thing will also be happening so it will be a very intense week as I already mentioned at the beginning but tomorrow you should be here again at 10 o'clock with the next academic session and we will have a birds of a feather, the doctoral consortium will also continue upstairs on the fourth floor uh yeah and we also invite you to stay with us at the cube here at the restaurant to reflect the first day and also talk to the speakers so I'll see you tomorrow. Thank you.