Diolch yn fawr iawn am wylio'r fideo. Thank you. Thank you. Thank you. Thank you. Teksting av Nicolai Winther Thank you. Okay, yes, now here we go. So welcome back to the next session. This is a session that is entitled Archive. There is a focus on cultural heritage, Living Archive at Ars Electronica. This edition on Saturday as well, so you are also cordially invited to join the Deep Space session on Saturday on the focus on living archive. But now we are focusing on archive from a research aspect. We have brilliant speakers, and the first one is Xiuwen Huang. She's a Taiwanese artist-researcher, interaction designer, educator, and founder of Vibe Weaving Studio, a creative researcher, practice bridging code, craft and cultural technology. Her work explores how generative system can support cultural continuity without flattering lineage, authorship, consent or community authority. So we are starting with the first talk with the title Lineage Canvas Productive Friction on Born Digital Heritage. Welcome! Hi, good morning everyone. So I'm Shiwen, you can call me Wen. I think I just gave a very brief introduction about my background. So this whole project is actually basically trying to, this is like a lineage canvas in the productive friction in Bon Jovi heritage. So this project is actually asking a very simple question. When situated material enters a generative system, how can source consent and restriction travel with it? The paper used foreign digital heritage as a case to propose a lineage-aware infrastructure for generative system because the plausible outcome can still savor relation. So generative AI change more than production speed. A form that once emerged through observation, apprenticeship, and engagement can now approximate in seconds. So the result may look like culturally plausible, but it's really hard to ask the question, so where is this come from? And who has authority to extend it? And also what obligation remain attached? So and the important, the most important thing is not only like imitation, it's also that the relation is actually being stripped away while the appearance may look very convincing. So the paper asks one question. What must a generative artifact remain accountable to? Three things. So the first one is the human who initiated it, and then the computational system that transforms it and also the source that make it possible. So consent credit and restriction shouldn't be expired at the moment when the output appears. Born to show heritage is the case. A lineage-aware infrastructure is the proposal. So in this prototype, so this is basically an artwork and also a research prototype. I used Taiwanese indigenous tribes as a use case, and because I have a background for weaving, so this is using three different tribes in Taiwan. The first one is Ataya. And so Ataya is basically using mapping. It's these three relationship basically trying to deliver mapping unequal. So one relation must proceed and the other is a proxy only. That one is the pi one and the other one is the refusal state, which is the ru chi. And the entire relation proceed with a condition attached. The pi one relation is also defined proxy and the root-key relation is encoded as a protected boundary. So recognition becomes a permission. So rather than just describe it, I can just demo this. So this is a research prototype. So when I selected Ataya, for the system, it's about you need to know the source and also how does this context come from. But for the human actually trying to drive this whole thing, you need to witness this and translate. And there's a governance state, and you're trying to understand the relations, and then there will be records for the consequence. And these are the record for each of the bone digital artifacts. And there's also another relation. So for RuKai here is being presented as a protected boundary. So recognition is not equal to permission. So the system is actually allowed to say no in terms of what happened. So this is also the lineage record. So like I mentioned, for RUCAI, it's actually being constrained. So refusal is not a failure. It is actually a generative outcome. When the protective root correlation is selected, recognition does not become permission. The system stops, but the attempt does not disappear. Source, restriction, and reason, and refusal remain in the provenance chain. So a lineage-away infrastructure must be able to produce a governed refusal, not only just an output. So I call this whole process like a productive friction. Intentional resistance, where the stimulus would erase responsibility. So you need to recognize, recognitions come first and governance comes second. Lineage confinement comes third. So whether the outcome is an artifact or a refusal stay, both persist with layer record. Friction is meaningful here, only tied to a condition. So this project makes three contributions. The first one is lineage-aware infrastructure, where the condition travels with the outcomes. Second, productive friction as a strategy, where delay or refusal become the mean for material, not a bug. Third, the heritage legibility. Can someone actually trace their origins and also identify the relations and also understanding what happened? And to be clear, this is an implemented prototype and R-based case study, not a valid scale or any community co-design. So there's no indigenous data set was being used for training, and no model inference happened at runtime. So I'll close here. Heritage makes the state visible, but this isn't only about heritage. It extends to any kind of generative systems working with situated material. Communities, archives, public memory, or data with social cultural stakes. The case is heritage. The proposal is infrastructure. Condition, travel, and refusal remains possible. And every outcome keeps a province record. Thank you. Thank you for the presentation. I would like to start with a first question. So born digital heritage is a question. So born digital heritage is a anything as being digitalized. So in this scenario, it's actually the weaving structure. So being like, it's another form, it's not actually being made, making like a tangible as a output, but it's actually persist in the digital world. And so, funny thing is I'm actually interaction designer and product designer, so for me to design a lot of software, I need to make sure everything becomes seamless. And so frictionless is actually the production, commercial company will ask for. But this project is asking another question. So because generative AI has become so fast, but not everything should be, you know, making so fast. Sometimes it's a pause to think about what's the state to generating something. Because that actually created a lot of problems. So people kind of like having cognitive overloading and they have no idea about what is the born digital artifact is being created. What is the source relationship and the context in terms of this items? So it's not just about the output. Sometimes the generative system, we probably can think about maybe they can allow to say no and refuse to generate as well because you did not finish the whole recognition process. Does that answer to your question? Thank you so much. Are there any questions from the audience? So if not now, at the end, there will also be the opportunity to address questions again. Here is one. For your practice, if you're in system design, or like just system design is not the proper thing, but how would you integrate friction to make your systems better or more interesting in that sense? If the conclusion, if there's something interesting or in that sense? If the conclusion, if there's something interesting in having that friction, where and why? And I guess in this context or society, how can you actually negotiate place for friction? So friction here in the context, in the system is not a bug. It's a moment to declare a state for people to pause, to think about what is the thing that I'm trying to do here. Because generated AI will give you everything. You just type in, using problem-based engineering and typing anything you want, and they will just give you the answers. But a lot of time, you are not actually thinking through everything. using problem-based engineering and typing anything you want, and they will just give you the answers. But a lot of time, you are not actually thinking through everything. You are accepting a lot of passive outcome. And people become like a human. This is like AI and a human relation. So human in this context is being... Humans should steer AI interface. But right now, it's more become like we become offloading to everything to generate AI. So that's why I think in terms of generative AI system, should have a stop mechanism. And everything should build into the infrastructure level. Not just like, for example, lineage canvas is creating the provenance chain that embedded into the infrastructure. So everything like being created, like a born digital artifact, anything digitalized being created, is actually embedded with the provenance chain, no matter is a refusal or actually a recognition being produced, it's all in the records. So everything you could trace what happened. I don't know if I actually answered your questions. Thank you. Thank you. Any other questions? So then let's move to the next presenter. Thank you so much. Thank you. Let's move to the next presenter. Thank you so much.. Let's try. Okay, I will. Yes, great. A very short introduction. Xiying Zhang is an artist and lecturer at the Taiyin Academy of Fine Arts, currently based in Beijing. She holds a PhD from the Central Academy of Fine Arts, Beijing, and studied in the Karlsruhe University of Arts and Design in Germany right now. Her research focuses on AI in artistic and design practice, working with fine-tuned diffusion models and generative image making, and has applied this approach to cultural heritage through AI-assisted restoration. So we're looking forward to your presentation. Okay, thank you. Hello, everyone. My name is Xin Zhang. I'm an artist and lecturer in digital media art at Tianjin Academy of Fine Arts in China. My work focuses on media art, internet culture, and the cultural heritage. I'm very happy to be here. I will introduce my paper and the artwork behind it. And the work is called Naran Shiv. So it is an open-ended interactive installation made of hand-built ceramic vessels. The surface design are based on painted pottery from the new SelSelic Magyar culture. Some of the vessels hold mobile devices. The screen shows the changing Python-like pseudocode and the short text based on the mess. And the viewers can speak, can type or clap or make some other noise. The system takes this action, but it does not reply directly. Instead, each action interrupts the historical simulation that is already running. So the work asks one question. Does the history have a hidden code? So, the simulation asks one question. Does the history has a hidden code? Or can the viewer's action become an unexpected event and change its direction? In NarrowShift, the audience interaction is now to command with direct response. Each action became an interruption. And this interruption built up inside the simulate history. So before I explain the technical system, let me begin with the title, the narrative of all shape of foods. The Boche's paintings use the shape as the of the humans' foolishness. The folk who described the shape of food as an image of exclusion, travel, and uncertainty. This influenced how I look at the early AI system. They maybe made mistakes, misunderstood people, and often produced foolishness results. For me, that foolishness the foolishness result. For me, that foolishness is became the start of creative journey. So if we want to talking about history, where does it begin? So I use the creation mass and the stories of culture heroes as a program starting point. Like as we can see, Pangu in Chinese traditional math separated the heaven and the earth. Like Sui and Shi taught people how to use fire, and the Pulusha comes from the Indian tradition, while the Eden and Eve from the Bible. Once this story are loaded, the simulation begins. So as we can see, the Pangu separated the heaven and the earth and the historical time begins. Here, I use an idea from the Reinhardt's co-science collection. He described the history time as a space between the experience and expectation. The system rewrites the creation stories from the difficult culture in the same Python-like language. In this program, the some cultural difference may become simplified or lost. That leads to my next question. What does this have to do with Ma Jiayao pottery? I'm especially interested in painters of the Chinese patient pottery. I also spent long time in Northwest China to making the ceramic vessels with local potters. Some scholars connect this motif with the early rituals and beliefs, although this idea is still debated. For example, some researchers connect the frog motif with fertility and the bird motif with the song. I recently saw a very similar thing, the pattern on the ceramic vessel in museum in Athens. This vessel comes from a different culture, but some of the visual form looks very similar. I treated that as a visual connection. In this work I combine these patterns with the coat-like signs and the symbol from today's inter-night culture. Through repetition and imitation the painted pottery carries patterns from one generation to the next. Both of the forms and the meanings are changed over. So this is the focus detail of my artwork. So they also remind us something important. Long before the computer appeared, the surface of objects could already hold the relationship, memories, the way of understanding the world. So the phone holders reading from the wasps looks like a snake from the out of them. So I chose the mobile phone because I wanted to, interaction to feel close and private. So as we can see the screen, the viewers need to, if the viewers want to see the screen, they need to walk up to the vessel and sometimes they need to bend down. So, now let me show how the historical machine works. A computer in exhibition runs local AI model by Ollama. Audience input becomes an interruption. The system updates the world state and generates new pseudocode for the phones. The world state keeps a record of the current mess and everything that has happened so far, including conflict and errors. So let's return to the Panguai example, at first, the heaven and earth has just separated. The system is still stable, and when the viewer says chaos fracture, the system save this word as an interruption and generated this code. The commander append scar, either the interruption to the timeline as a scar. So the interruption to the timeline of the scar. So the scar stay in the word state and affects what happen next. Please look at the last line of the code. The year equals minus one. This is observed result is intentional. The system has lots of normal position in time, but it continues to generate history. I describe this process with three words, interrupt, patch, and drift. Each new interruption pushes the system farther from the original path. The movement is a drift. So the work is also related to glitch aesthetics. During the text, I noticed the interesting change in the way the people behave. At first, they may be a little nervous or maybe pilot. They spoke in full sentence as if they were talking with a chat bot. And then the system did not reply, and some began to cry or laugh or make awkward noise. And then the viewer focused the changes. They began to ask, where did my action live or trees? The viewers can affect the system, but they cannot control the result. To me, this is very similar to number place in history. So let me back to the question, does the history have a hidden code? After making this word, I still have a clear answer. So the narrative is a small experiment. Each run carries three traces from the one before. The work also changed now how I understand the title, Nero's Ship, or Ship of the Foes. The ship has no fixed destination. It could be arrow, it could be noise, and the scar travel with it and keep changing it along the way. So the work makes this uncertainty easier to feel and the question remains open. And that's all my presentation. Thank you very much for listening and looking forward for the connection. Thank you. Thank you for the presentation. I'll start with the first question. I'm interested in the conversation with art historians and experts in the archive. So when I had the privilege to work on a cultural heritage project, we observed that the experts on the archive and cultural heritage artifacts, they gave us different answers and our questions started a discussion about specific questions that they have not considered. So my question is, was there a dialogue with experts on the cultural artifacts, and how was the dialogue, the conversation with them? I think if I'm understanding correctly, you want to talk about the dialogue in the history? Yeah, yeah. It was exhibited, so it can be, of course, a conversation in the pre-production, but it could also be at the exhibition when experts with expertise in the cultural heritage objects, what was the dialogue? expertise in the cultural heritage objects. What was the dialogue? Have they reflected, discussed the project? Yes, for this project, maybe in the past year, I have research on the how to restoration the ancient pottery, how to restoration the ancient pottery, how to restoration the patterns with the AI tool. And in this project, I'm already focused on the interactive between the viewers and the program. So maybe in the common interaction installation, the viewer wants to get a certain reflect. And in my program, I'm not engaged in certain results because I think, like in my program, I was talking about the history, like not a very smooth line. So I think it's a totally different thing. So for the cultural heritage, I'm focused on what is exactly how could I do with AI to restore the patterns. And in this program, I'm very engaged in how the people to see the interactive is valid, or can solve any problem. So I think it's a totally different thing. Thank you. I think the conversation from an artist perspective, and art historians, or experts in archive, I think that's very fruitful and to get their new insight in terms of the knowledge that we have in the archive. aspect but the only thing I'm very focused on is how the people understand the world, so how the technology to change our world of the change our mind of the world or change our mind of the history and the history is maybe we can say we cannot affect history actually, but we can leave a scar in history. Maybe that is the meaningful life for every person to do. Thank you. We have time for one short question. Thank you so much. Thank you so much. We will continue with the next presentation. So it's Dayan Wang, right? Welcome on stage. Dayan Wang is a PhD candidate at the Queensland University of Technology. His practice-based research explores how extended reality, immersive environments, and digital storytelling can support cultural reconnection with diasporic communities. diasporic communities. His current project, Root Seeking, investigates the reinterpretation of Chinese cultural heritage through our experiences. Welcome. Hello everyone, my name is Daichun Wang and this project was completed with my supervisor Dr. Rehwald, Professor Damian Conduzzo, and Professor Gavin Seed at Queensland University of Technology. And today, I will present our artwork, Roots Thinking, or Xun Gen, which means searching for where you come from. And here is referring to reconnecting with someone's personal history, place, and cultural memories. So what you see here is not a photograph, it's our reconstruction of a traditional Chinese garden inside VRChat. First, let me introduce what the work is about. The artwork was built in Unity 2022.3 for MetaQuest 3. It engages users through the gesture, sound, and poetic narratives. It engages users through the gestures, sound, and poetic narratives. Its visual language and spatial structure come from Jichang Garden, which is a Chinese garden from Ming Dynasty located in Wuxi. And the garden mainly shapes how the virtual space is experienced, through the frame views, board sensory, and the view that gradually appeared. through the frame views, board sensory, and the view that gradually appeared. But our goal is different from some XR heritage research, which often focuses on the accuracy, reconstruction, education, and tourism. This work is made for different audiences, people living far away from home, which we call it here diaspora. Instead of not instead of only recording the garden, it used the garden as a way to reconnect with the culture and memory. This also means giving the diaspora more voice, but this creates a limit as well. The research team still decides what to scan, what dialogue to use in the virtual environment, and which garden should stand for home. And one, Jichang Garden, which is what we use as inspiration, cannot fully represent the Chinese heritage. So that leads to an important question. the object is not a real object, but it is a real object. So, we can see that the object is not a real object, but it is a real object. So, we can see that the object is not a real object, but it is a real object. So, we can see that the object is not a real object, but it is a real object. So, we can see that the object is not a real object,. Below is a fact layer. We have 1.2 objects here, systematically scanned and published openly. In the archive on the right, every object here has its own page, so anyone can check and can record it. A void is a supplicated layer. Ritual, narrative, and atmosphere create with the community and clearly present the interpretation. So, participants are afraid to imagine, but that imagine does not replace the historical record. And let's jump to the... Yeah. historical record. And let's jump to the, yeah. And let's first show you what we mean by the fact layer. We capture 102 objects and grounded them into six categories based on their role in the garden. The largest group is architectural element with 46 objects. And this is partly because those objects were easily to reach and record. While the smallest ground were much harder to capture. So we used three main methods to capture those objects. The first one is iPhone LIDAR, was used for the object at ground level. Gussing and splitting was used for the object at ground level. Gussing and splatting was used for the detailed surface, and drawing photogrammetry was used for larger surrounding views. Vegetation was much harder to capture, so all the trees we built in the environment was created digitally. Gussing and splat also become more than just a technical method because it made from the point and light rather than just solid surface. It gave us a visual quality and that feels close to memory. Clear in some space but less solid and completed from the distance. Let's jump back a little bit. But building the garden was not only one part of the project. We also needed to understand what people actually want from a virtual heritage space. This slide shows one of the most important funding in this project. We also needed to understand what people actually want from a virtual heritage space. This slide shows one of the most important funding in this project. The footnote shows where the data came from. An on-site survey, an online Diaspora survey and an on-site prototype demo. The chart comes from the on-site study. Visitors gave Chinese gardens a high score of cultural importance at 4.47, but the score for the personal meaning was low at 3.53. So people could see the garden as of cultural importance without feeling strongly personal connecting to it. We saw a similar pattern in the emotional attachment. Visitor score 3.80, where the staff, the staff is the garden administration. In the first on-site questionnaire, scored like 4.5. Participant with a more long relationship to the garden were reported stronger emotional attachment. But this is a small sample. I treat this as an early observation rather than a casual finding. But it's important for evolution. People may recognize a virtual heritage space without feeling attached to it. On the right, we can see what this means for design. Sensory experience was rated as the most important feature, and at the same time, 14.1% of the participants had never used VR before. And the clearest result was that people wanted to avoid this 130 out of 172 choose overall complex control. Those finding directly shaped our design choice. Participant want a strong sensory experience, but they also want the simple interaction. So instead of building only one type of virtual space, we explore two different way of responding to their needs. So we built two different worlds, and I would like to show you both. The VR chat world is more than just being inside of the space. It uses continuous environment, realistic scale, sound from different directions, and a complete visual scene. It encourages participants to look around and experience the garden. The Minecraft world takes a different approach. It's more than just making and changing things. The block system is simple, easy to understand, and easy to rebuild. It makes the world feel open to change. In our demonstration, we tested the project as one overall experience. So this comparison is mainly based on what each platform allows people to do, together with the survey funding that people prefer simple interaction. And I will first focus on the VRChat version and how people experience it. Up to three people can use the wall together on Quest 3. And they can move freely in a 3 by 3 meter space, running and the Quest 3 is running at 90 FPS. There's also a desktop version for people without VR, so they can get access to it through their computer. When users move close to the artifact, they can hear the classic poetry in Mandarin and local dialects. And there is also a Sujo Pintan and other dialect sounds are placed in different part of the garden, so the sound feels connected to the space.. And the second platform, Minecraft, has a different role. The Minecraft world runs on the Java edition with our own server. I also include the AI cultural guide through the service called MyPair. The guide can talk with the user about Chinese garden, classic poetry, and Chinese philosophy. It can also change the language when the user asks. So one point is important here. We configured the AI rather than just building the AI system ourselves. We draw the guides, characters, and instructions based on the selected cultural resources. The AI follows the construction and the instruction during the conversation, but we do not treat the answer 100% correct during this experience. Instead, the guide works as a exploration, cultural interface with the co-design process. The Minecraft board also includes message board, voice chat, and the personal county art. And the participant can change and build all the personal county art. And the participant can change and build all the environment by themselves. So Minecraft world mainly as a co-design and workshop space, rather than just a world to explore. The co-design workshop will continue through the second half of 2026. Actually it's already happened. And with Minecraft serving as the primary workshop environment. And together, those two platforms lead to three main contributions. First, from the participant data, we learned what the audience values. They want the simple instruction rather than game-like tasks, change over time, like the day and night cycle, richer sensory content and simple way to use the system. Second, the design research and the main contribution is the two platform model. The VR chat provides more details and immersive space. While the Minecraft provides a simple space for making and co-design. The two are connected as part of the same design process. Third, this is an art practice. The project explores how the cultural heritage in XR can focus on feeling, memory, and visual, rather than just virtual realism. There is also an important limit. So far, the funding shows us user needs and design preference. They do not yet show the long-term effect. Question around the identity will be examined after the workshop. More broadly, this project suggests that heritage place can be shared virtual space for people living far away from home. Where the cultural memory can be experienced, changed and rebuilt together, the archive is also available on the Sketchfab if you just type Ruth Higgin as the keyword. And that's all about my presentation today. Thank you guys for listening. Thank you so much. Thank you for coming from Australia to Austria. It was a long journey. Greetings to your co-authors. They could unfortunately not make it to Linz. So the same question, have you also talked to experts and how they gave you feedback especially in terms of realism versus abstraction in terms of reconstruction of cultural heritage? Actually I have talked to a lot of expert in this area, but the expert I would like to share today is a Chinese grandfather. I met him in the Jichang Garden. He basically just walked me through every construction of every important fact of the garden, and he taught me like, for example, this is a window, but the video can be treated as a frame if you are photographer the frame would be the most important element if you're doing photographer the film can be treated and through the frame you will see another landscape but after frame the landscape just changes so basically our design followed more to the grandfather scene rather than just like some expert in the 3D. That's the most important part we found in this project and it happened in our first group. The participants loved this kind of design rather than just we have a really high level like 3D modeling and stuff. Yeah, they like this kind of design we find in the real garden rather than just from the expert. I observed that if the experts don't have a clear answer, it's more like, oh no, not realistic visualization or details because we don't know. Is there a second stage in reconstruction of a historical building or not? Then say, oh no, no, a more abstract version. So that's my observation. Any further question to the project? Victoria. Thanks for your presentation. I was really interested in the ways in which you aligned the different platforms with different kinds of strategies. So the VR chat with more of the presence and a more built out kind of environment versus the Minecraft blocky state. And it leads me to wonder where you think the future is for this kind of work. Because right now it seems like you have to have two different work paths perhaps. And then we all know that these platforms go away as soon as we get everything all built up nicely. So could you say more about how you're thinking about how to privilege development in the future for this kind of work? For me, probably I would like to run both platforms, because from our survey, most of people who like to take our survey, they are from Gen Z. And Gen Z, and based on our literature review, Gen Z are more close to the Minecraft rather than the VRChat. Because VRChat just I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. I think that's the thing that I think is really important. previous slide. They would like to change. They would like to build their own apartment. They would like to change the garden. In that way, they can really experience and to find their own identity when they participate in the garden. So in that case, in order to improve the diversity of our project in the future, I would like to run both two platforms and keep them running alive. Any further questions? Thank you so much for your presentation. Thank you. We will continue with Emily. Emily DayDam works as an academic artist and creative technologies in the space of immersive storytelling and social justice. The stage is yours. Thank you. Hi, my name is Emily. My pronouns are they, them. And this research builds ar fy PhD, a dwi wedi'i ddefnyddio yn gyntaf yn gyfranol yn Fawrth. Mae'r bwydau bwydau bwydau i mi yn dynnu, yw duned, re-exar, heredigaeth wir, obsolesn, exar, cyn ddiweddio'r hoff. Peidiwch â'i ofyn, os nad ydych chi'n gwybod beth mae'r rhain yn ei olygu. Byddwn i'n ymwneud â'r rhain. Byddwn i'n dechrau gan ddweud bod fy nghymorth yn ddrwg, ac roedd yn rhaid i'r rhan fwyaf o'r PhD ddeall hynny. Mae'n fflavwr arbennig o ddrwg, a dyna'r ddrwg e-gwaest. Mae e-gwaest yn ddrwg eledronol. Mae'n ymwneud â defnyddio ariannol ariannol sydd wedi cyrraedd y diwedd eu bywyd ddefnyddiol. Mae'n un o'r drwg e-gwaest fwyaf yn y plannad, gyda 347 miliwn o ton o ddwyloedd e-bwyd wedi'u cymryd i fod ar blynyddoedd yr Eiriaeth ei hun, ac mae hynny'n ddiweddar y flwyddyn gwaith ystod y llaw. Gallwn ni ddim ystyried, gyda llawer o statiau, ond mae 17.4% o ddwyloedd e-bwyd yn cael ei gael cael ei gael a'i gael i'w gael yn iawn. Dyna dim ond y cymaint o ddwyloedd e-bwyd. Rwy'n credu bod y sector ymersif yn cynyddu'r cyfrifoedd e-bwyd yn gyflym. Gwnewch Rhe XR. Mae Rh e yn ymwneud â rhwydweithio ac XR yn ymwneud â phwerthu. Rhe XR yn ddefnyddio cyfrifoedd sydd wedi cael eu llwyddo gan y sector XR, ond gall eu rhwydweithio i ddweud y storïau sydd wedi cael eu rhannu neu ddim yn cael eu cyflwyno mewn archifau. Mae Rh XR yn defnyddio dechnolegau golli i ddweud storïau golli. Y pwysig nesaf yw cerddedig wirthu. Mae deffiniad Sarah Kenderdean yn y cerddedig rwy'n ei ddefnyddio yma, nid yn unig oherwydd roedd nhw'n un o fy myfyrwyr PhD. Y defnydd o gerddedig wirthu a gwahaniaethau gwahanol mewn ymchwil cerddedig diwylliannus i weithio ar y cyfran rhwng ymchwil gwyddonol cynnyrch a'r datasynau a deitlau datasyn cyfryng a'ol ar gyfer y cyfranogol ymchwil ar gyfer y cyfranogol ymchwil ar gyfer y cyfranogol ymchwil ar gyfer y cyfranogol ymchwil ar gyfer y cyfranogol ymchwil ar gyfer y cyfranogol ymchwil. Mae'r ymchwil hwn yn canolbwyntio ar sut y gallwn ddatblygu ymchwil ymchwil ymchwil yn ddiogel. Mae'n werth nodi bod llawer o ddefinidiaethau o ymchwil ymchwil a therbyniadau eraill fel ymchwil digidol. Roeddwn i'n hapus i glywed pan oedd ymchwil digidol wediol hanesu, archeoleg arsgolion cyfoed ac llawer mwy. Efallai bod yna hyd yn oed ffordd i gael hanes XR. Felly, pam Dunedge? Dwi'n ddigon ddechrau gyda hyn. Mae'n ymwneud â'r gwaith gwaith i'w ymchwil. Dwi wedi dewis y ddyned o Dunedge fel sefydliad i'w ymchwil oherwydd mae ei hanes wedi cael ei ddysgu drwy amser, gan ystyried sut mae amser yn ddysgu technolegau XR trwy'r cyfnod oesol. Yn unig y byddai'r dyned wedi cael'i llwyddo i'r môr Cymru, mae'r ddwylliau digidol wedi'i llwyddo i'r rwst digidol. Mae'n gwneud, o fy marn i, safbwynt arall i ysbrydoli methodoleg arall arall. Fel rhywun sy'n tyfu i fyny yng Nghymru, rwyf wedi gweld y newid ystod y llaw, ac fe wnes i'r cymhwyso i gynhyrchu ac adeiladu'r ardal hwnnw o her i'r cyfnod o ddiwedd. Roedd yn cael ei ddod o hyd i'r cyfnod o ddiwedd. Roedd yn cael ei ddod o hyd i'r cyfnod o ddiwedd. Roedd yn cael ei ddod o hyd i'r cyfnod o ddiwedd. Roedd yn cael ei ddod o hyd i'r cyfnod o ddiwedd. Roedd yn cael ei ddod o hyd i'r cyfnod o ddiwedd. Roedd yn cael ei ddod o hyd i'r cyfnod o ddiwedd. Roedd yn cael ei ddod o hyd i'r cyfnod o ddiwedd. Roedd yn cael ei ddod o hyd i'r cyfnhau o'r llinell ymddygiadol yma yn ddi-wyd. I'r rhai sydd ddim yn ymwybodol â Dunedge, mae'n enwog fel Atlantis, Brifysgol. Felly, mae fy ngwestiwn yw, sut all dechnolegau creadigol ddatblygu hanes wirfoddol yn ddiogel pan fydd y gwaith maen nhw'n ei ddatblygu'n ysgwyl yn gyflym â'r amgylchiadau maen nhw'n ceisio'u diogelu? Fel datblyguoedd yn tech ymddygiadol, rydyn ni'n adeiladu ein casgau ar ysgol. Gyda'r cyfnodau sy'n newid ymhellach i'n gwybod pan fydd ein casgau'n llwyr. Mae'r ymchwil hwn yn golygu i ddatblyguwyr gweithio mewn gweithdai gynhyrchol. Os na allwn ni fod yn ddatblyguwyr gynhyrchol, sut y gallwn ni ddatblygu profiadau XR yn awyddus? Yn ystod y rhôs, mae'r rhôs Dunedge yn stori gwrthdaro XR ar gyfer safonau ar gyfer ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod ystod gwyddoniaeth ar-lein ar gyfer ystod ystod y llawr. Mae'r dynfod yn ddysgu'r stori gwyddoniaeth ar-lein ar gyfer ystod y llawr. Mae'r dynfod yn ddysgu'r stori gwyddoniaeth ar gyfer ystod y llawr. Mae'r dynfod yn ddysgu'r stori gwyddoniaeth ar gyfer ystod y llawr. Mae'r dynfod yn ddysgu'r storiryd enw Rose i ymwneud â sgwrs dynol. Mae'r stori yn dweud eu bod nhw'n gofio pan oes eu tŷ a'u teulu wedi cael eu llwyr o amgylch ymlaen. Mae Rose yn dyn a'i stori yn ffyrdd o unigol. Rwy'n gweithio gyda dynion i ofyn y material a chynaliad o amrywiaeth stori. Mae'n ddiddordeb i'r maes o stori. Mae'r heriau wedi'u cyflwyno. Yn ymchwil, fe wnes i ddatblygu'r rhaglen ar gyfer defnyddio methodau modul, agored a thafodd. Mae'r rhaglen yn lles arferol. Er bod y gwaith yn arbennig ar gyfer y lle, gallai'r gwaith ei ddatrys i leiaf. Gallwn ddysgu sut gallwn ddweud stori'n ddiogel.....yn ffwrdd agored yn bwysig.....gyda gwasanaethau cynllunio, gwaith ddewis a chyflwyno. Yn ymwneud â heriau, roedd yna'r cyfnod o ddiffyg.....o dechnolegau fel ystafelloedd haptig.....a'r rhain yn aml gynnal asthedig llyfn neu'r arfordir. Dyma ddau enghraifft o ddigwyddiadau cyfrifol. Mae fy nghymorth fatholegol yn un sy'n arwain ar y pradwyth. Mae'n un cymysgedd. Byddaf yn dechrau gyda media zombi. Nid oedd unrhyw zombi yn cael ei ddiffyg yn ymwneud â'r ymchwil. Mae'n ymwneud â media sy'n peidio â thamgymryd o'r cyflenwad planetol. Yn aml, mae'n ymwneod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r cyfnod o'r ffabwlaeth cridigol a'r meddwl. Mae'r ddau ddau ar y dde, yn dangos ymgyrch diagraffu y gwrth-gyrch Gartner Hype gyda media dynol. Mae'r cyrch Gartner Hype yn ymgyrchu â'r ffyrdd a'i ddangos pan nad yw'r cynnyrch yn gwneud cyflwyniad cyflwyniad. Mae Re-XR yn byw yn y ffazen dynol. Mae'r media wedi ffail i'w wneud i gyflwyniad cyflwyniad neu wedi dod yn ddiflannol neu wedi cael ei ddysgu. Rwyf hefyd yn gydgyddeol ac yn debygol o fod yn rhaid i'r cyfrif gael diagraffu cymhleth iawn ar ryw bwynt arr cyfnod hon yn ddiddordeb, ac mae'r ffyrdd oedd yn ddiddordeb. Mae'r ffyrdd oedd yn ddiddordeb, ac mae'r ffyrdd oedd yn ddiddordeb. Mae'r ffyrdd oedd yn ddiddordeb, ac mae'r ffyrdd oedd yn ddiddordeb. Mae'r ffyrdd oedd yn ddiddordeb, ac mae'r ffyrdd oedd yn ddiddordeb. Mae'r ffyrdd oedd yn ddiddordeb, ac mae'r ffyrdd oedd yn ddiddordeb. Mae'r ffyrdd oedd yn ddiddordeb, ac mae'r ffyrdd oedd yn ddiddordeb. Mae'r ffyrdd own yn ddysgu'r grip dynol ar archifau hanesyddol, yn gofyn am ystafell ar gyfer clywedau hir a'u gwneud yn fawr i'r dyfodol trwy ddynol ymddygiadau radigol. Dyma sleid ffotograffig o'r ffordd mae Rose Dunwich yn edrych. Yn rhan o'r ystyried y gwaith hwn, mae'n ddewis nad yw'n Dunwich Beach, ond brofiad arbennig. Yn rhan o'r cyfathrebu, fe ydym yn gofyn y cwestiwn nesaf. Beth oedd y rhan fwyaf effeithiol o'r profiad? Roedd pob ysgrifennydd yn dweud bod y fest haptig yn alluogi eu synnwyr ymddygiad. Roedd y cyfathrebu tactil yn galluogi defnyddwyr i ddiddordeb yn ffusigol i'r sain y môr, yn dangos y gallai gwaith arferol ffostro ar gyfer cymysgedd sensoriol, ac yn rhywle i'w hwynebuwaith o'r systemau gweithredol. Nid oedd unrhyw un yn sylweddoli bod y profiad wedi'i datblygu yn unigol o ddefnyddio media gwasanaethol, yn annibynnol o gyfrifoldeb sy'n meddwl pam nad oedd gen i set o glasau enriol o'r ffordd o'r ffordd o'r ffordd. Gwasanaeth, gynnyrch arall. Yn ystod y cyd-dysgu, mae'r AI yn ffordd arall o ddysgu technolegol, gyda cyfrifoldyd dechnolegol, gyda cynydd o obsolesiwn technolegol fwy na rhaid i'r cynydd o ReaxR oherwydd cynydd o leithiau bywydau cynhyrchion. Byddai AI genedlaethol 2030 yn gallu cynnig 5 miliwn o tonnau metric o ddwyloedd. Mae obsolesiwn yn dod i ni i gyd, ac os nad yw'n dod i chi, gallaf ystyried nad ydych chi'n ddyn iPhone. Mae Kate Crawford, sydd â cael atlases o AI yn y tu ôl, yn argyfwng bod, wrth i'r capabiliadau AI ymlaen yn gyflym, maen nhw'n gadael cymaint o gwaith o'r cyd-dynol, gan gynhyrchu lles o gwarthegau sy'n gweithio'n ddifrifol a oedd wedi'i fframio fel cynnyrch technolegol o ddiwedd. Mae'r ymchwil hwn yn cyd-dynol a'i gy chyflwyno'r angen i de-carbonio ymarfer amrywiol wrth gyfrifo'r criseu o'r hinsawdd a'r diogelu cymdeithasol. Mae'r sleid hwn o 1911. Fe wnes i ymwneud â'r ddelwedd hwn oherwydd roeddwn i'n teimlo'n debyg o'r ddau tref yma yn ystod y phrifysgol. Rwy'n cael fy ngwrdd â'r ddilemau rhwng y ddau. Mae yna gynyrchiad mewn nifer o gyfresoedd o ddysgu sut mae XR yn cael ei ddefnyddio mewn stori-tellyn amgylcheddol. Ond mae'r astudiaethau hyn yn fawr yn canolbwyntio ar sut y gall y ddewision hynny gynyddu llythyrwyr gynaliadwy. Nid wyf yn ddiolch ar y syniad hwn, mae hynny'n wych. Ond rwy'n credu y dylem ni ystyried effaith amgylcheddol y gwaith gwaith a'r cynyddion bywyd a chynnydd a chynnydd ariannol yn ymwneud â hyn. Mae fy nghymorth i ddatblygu lles yn cynnwys cymryd ymddygiad cyflawnol i ddatblygu profiadau XR a'u creu drwy ddefnyddio e-waste a chyfeirio llwythoedd agored i rannu profiadau a all eu ddysgu mewn brwswr weithredol i alluogi mynediad i'r hanes wirfoddol drwy gyd-dysgu ac ddefnyddio cyfnod o ddigidol cyfnodol. i'r hanes ymddygiadol drwy'r cyfrifiad ac yn defnyddio'r gwaith cyhoeddus digidol. Mae'r ffyniad hwn yn dangos sganner LIDAR gwasanaethol. Mae'n sganner mewn goffi, ac o'r cychwyn i ddechrau ei ddatblygu yn y pandemig gyda'r hyn oeddwn yn ymgynghori o ran fy nghyfarchiad llythyrol, roeddwn i'n ei wneud o'r cychwyn i ysbrydoli sut i ddatblygu defnaithau ganolbwyntio ac hefyd oherwydd nad yw fy nghyfarwyddwr yn ei gadael i fy nghymryd, neu eu bod yn meddwl y byddaf yn gynnal eu caner £20,000. Mae yna rywbeth da. Mae yna rhai strategaethau sy'n gallu cael eu ddefnyddio mewn llyfn yn erbyn methodau cynhyrchu. Mae codi gwydr yn un o'r enghraifftau. Mae'n defnyddio ymarfer cyfrifiadol i leihau'r arian sy'n ymwneud â'r broses cyfrifiadu codi. Mae hefyd wedi bod ymchwil am effaith y sefydliadna hefyd astudiaethau ar y pwynt o'r ddisblawn.....a'r cyfrif bwyd. Dyna pam rwyf wedi cyflawni'r llaw y gwnaethom ei gyflawni. Mae gen i asturisg ar y cyflawn carbon ar y dde. Mae'n debygol o fod yn cael ei ddefnyddio fel gwysio gwyddo. Ond nid yw hynny'n bob amser. Os ydych chi'n edrych ar gyflawn carbon.....rydych chi'n gwybod bod y cyflawn carbon yn gweith yn gweithiol, os yw unrhyw un yn edrych. Rwyf wedi dyfodol y byddaf yn ddiweddar, ond rwy'n gwerthfawrogi nad yw hyn yw'r ddyniad mwyaf gobeithio. Roedd yna dyniad o'r dyniad yn ystod ystod y wythnosau yn y flwyddyn, ac fe wnaeth gael 150 oed o'r dynion. Mae fy nheulu yn ddiweddar yn wythnosau ar ôl y llifogydd hwn. Mae rhai rhai, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyr, bydwyrch ar y sgriniau glwyddo sy'n rheoli ein bywydau a'n ein hwylio. Rhaid i ni ystyried y tol cyffredinol a chyffredinol ein bywydau digidol. Mae'r ymchwil hwn yn ei ddodd i'r cyfrifon dynol mwy na phobl sydd wedi cael eu cyflwyno gan y Gwrdd Gwyddo a'r Cwrdd Cymru yn gychwyn. Mae gen i list o gyfeiriadau i unrhyw un sy'n edrych. Rwy'n hoffi ddweud diolch. Diolch. Diolch am y cyflwyniad gwych. list of references for anyone curious and just want to say thank you. Thank you. Thank you for this great presentation. As we mentioned at the beginning there will be in a couple of weeks proceedings and then you can read the paper in detail. Thank you so much. I have a question, a practical question. As you are now an expert in terms of sustainability, digital born artifacts archived. So at Ars Electronica we have a very small physical archive downstairs in the basement. Books, brochures, but nothing in terms of artifacts. So the focus is, since many, many years, a digital archive where we have a lot of documentations of the projects that are exhibited at Ars Electronica. So the documentation is very, very important and also the description, a tech writer. So if someone wants to do it again 20 years later, there should be of course a detailed description to install the installation and of course you need the hardware as you mentioned. Someone has to take care that there is still the hardware from the 90s so that you can run it. So that's very, very tricky in terms of sustainability. So as far as I know, As Electronica does not collect hardware. Any advice for such challenging task to archive digital-borne media art? Oh, that's an amazing question. I think we're... So different heritage institutions and museums o'r cyfnod cyffredinol? Mae'n gwestiwn anhygoel. Mae amgueddfeydd gwahanol o ran sefydliadau a chyfnodau o ran amser yn cymryd rhan o'r mater hwn. Mae Tate wedi cyhoeddi nifer o ddau papur ar hynny. Yn fwyaf, mae pobl yn ceisio casglu cymaint o'r gwaith ffisigol neu ddefnyddio emulatio. Mae'n ffyrdd o waith ar wahanol elfennau o ran y gwaith, oherwydd nad yw un elfen ar ei hun yn gweithio. Mae casglu CRTs i gyrru llawer o ddysgu fideo o'r 80au a'r 90au ddim yn gwirioneddol, ac mae emuladu ddim yn y ffordd gorau. Un o'r ymarferion gorau ydw i, yw'r term Paradata, sy'n dod o'r ymdrin o'r hanes. Mae Paradata yn ysgol sy' y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Mae parodata yn ystod y maes o'r heredigaeth. Felly, byddai'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dweud, mae'n dwe of things that will evolve along with the technologies evolving. At the minute, we're kind of on the back end as kind of consumers developing. And I think we should try and lobby to reverse that so that these technologies are less proprietary and that we have more agency over them. So then we can control the futures of the works that we're developing a lot easier. Thank you. Any further question? Victoria. Thank you. Any further question? Victoria. Thanks for your presentation. As you were speaking, I was thinking about our desire to keep everything, my own desire to keep everything, and the fact that my house is bursting. But also, going to Juergen's question, do we need to archive everything? We have this back it all up, copy it, make sure it's instanced in the most rigorous way possible, and yet the history of media is also the history of loss. Should we just let things go? I'd be inclined to agree. We do have a massive... Perhaps if you're an AI, you'd have a different perspective because there is a thirst for data, and data is knowledge that's going to fuel a lot of these different models that we're using. Efallai, os oes gennych chi AI, bydd gennych chi safbwynt gwahanol, oherwydd mae yna ddyniaeth am ddata, a'r ddata yw'r gwybodaeth sy'n mynd i ddod â llawer o'r modelau gwahanol y mae gennym ni. Yn ffordd bach iawn, rwy'n eich herio i chi edrych ar y data y byddwch yn ei ddefnyddio yn eich bywyd bob dydd. Rwy'n credu bod 0.025g yw'r cymaint o emisiynnau CO2 sy'n dod o un e-bost heb ysgrifennu yn eich e-bost. Felly, gosodwch eich e-bost heb ei ddod o'r e-bost. Felly, gwblhau eich e-bostau os ydych chi'n gallu. Byddwn yn dweud na i... Mae llawer o amgueddfeydd, ac ati, felly rwy'n gweithio yn fwyaf gyda amgueddfeydd, ac ati, felly byddaf yn mynd i fod yn ddibysig ac yn ymwneud â nhw, ond mae colledwyr eraill ar gael. Mae yna ymddiriedaeth i ddigidu popeth a chyhoed-dau. Mae yna ddyn o'r enw... Dwi wedi gofio'r enw, ond gallaf ei ddod o hyd i. Roeddwn i'n dod o hyd i dyn o'r enw Palliative Care. Mae'r dynion yn ein lleoliadau. Yn ystod Dunedg, mae llawer o bethau wedi cael eu llwyddo. Mae'n cydnabod nad ydym yn gallu chadw popeth. Beth yw'r peth mwyaf pwysig? Sut ydym yn gallu chadw pethau sydd yn anghyfansodol? Dwi'n dweud na. we're not going to be able to protect everything. So what is most important and how do we protect things that are intangible? So yeah, I'd say no. I love the palliative care idea. If you had a question, we have time for one question. Thank you so much. We will continue to talk about archiving later. Thank you. So our last speaker is Xinran Wu. Are you here? Please creative strategist based in the Bay Area, working at the intersection of cultural memory and spatial practice and generative AI. You are talking about a project Anywhere Doors. The stage is yours. Thank you. Hi, everyone. My name is Xinran, and I am an independent artist and also a designer and creative strategist in AI industry. And it's my honor today to talk here to share my project, Anywhere Door, Tethered Hallucination as a Media of Remembering Shanghai, Vanishing Longtown Alleys. And it is a research on history, memory, and methodology. And that's the part I'm going to focus on today. And also that's installation or in progress. That's the part I'm going to focus on today. And also that's installation work in progress. So let's begin with imagination. Imagine you are an anthropologist from another planet and arriving in Shanghai for the first time. Local people tell you that there is a kind of building called Nongtang, which is a nice place to look at. that there is a kind of building called Nong Tang, which is a nice place to look at. And you see the three corners like this across the metropolis. And in outdoor, there is one type, like a shopping street, where renovated facade enclose luxury retailers. And second, there is a nostalgia-style food court staging slavage props taken into into the basement newly built shopping malls. And third, there is also a set apart from the prosperity. There are depopulated ghost towns awaiting for demolishing. Each of the places claim to preserve the memory of Nongtang, which makes you really confused why it looks totally different but calls the same name. However, each of them, in my opinion, is missing something in two ways. So first, the history has layers. Nongtang developed since the opening port of Shanghai in the 19th century, is in that time the newest housing type with a hybrid of Chinese and Western forms built for wealthy families. Through the wars and the changes of land ownership in the mid-20th century, it became dense collective housing, once shared by even 72 families on one roof. And third, in 21st century, the force of urban gentrification in the heart of Shanghai pushed it to restore it. And when one possibility right now is being built and fixed in break, the memory of every other layer is being worn away. And in a second, the Nong town is a space of everyday living, and individual lives have no mark that a structure can carry. Nong town once housed 70% of the city's population in 20th century, house 70% of the city's population in 20th century, and not as a heritage, but as a home. It is dense, sheer, improvised residential space where every life accumulates on every surface. As Henri Lefebvre would call it, a spatial memory. And the three places you saw are imagination of non-town memory onto the surviving architecture structure right now. Yet, none of them preserve this daily spatial practice at all. So, seeing these questions, where does the memory of Nongtang really live if it's not in the structure? And let us go back into the field. And from 2021 to 2024, I sometime kept returning to the ghost town version of Nong Town nowadays before it's being demolished. And there is an afternoon I stood in front of the door like this. On its surface, there is a milk delivery box with the household name in blue ink, and the spring couplet with red had faded to the color of dyers, and a lockbox of parcel, an address plate from one administrative area bolted above an address place from another. A door is not only an architecture element, it's also a threshold. Standing outside, I try to imagine what is the life behind this door. As Susan Scoopley argued in Material Witness, objects do not merely document events. Through weird, fading attachment, they testify carrying political, emotional, and ethnical weight in the proportion to their apparent insignificance. So there are the fragments I collected from that ghost town version of Nong Town, from Henry's store, and sorted them from the historical evidence behind them. You can see it's from different time eras, you can see also there are different occupations of the owners, and also there are different functions of the room, like the time maybe from 80s, 90s, early 2000s, and even pandemic. And from all of that, an imagination of life inside can be formed. And with this concept, the process in human brains, like collecting the pieces and decoding them, we may call it as memory. And there is maybe a parallel process in artificial intelligence. So memory, as Bartlett showed, is not a recording mechanism. It's an active process of reconstruction in which each recollection reorganizes available fragments according to present expectations, filling gaps between plausible materials to produce a coherent account. The past is now stored. It is reassembled each time when it is accessed. Each memory is both faithful and fabricated. Well, in a language model, maybe it does the same things. It takes fragments from the database and fills the gaps with what we may call reasoning. And when the fragment relatively is real and the whole narration maybe is not unconfirmed. We call it hallucination. So, hallucination and remembering, I'm not arguing this is a claim about their technical architecture is the same. It is an observation about a sheer epistemic pre-decimate. In both cases, a coherent account of the past is assembled from the fragment in the absence of the original by a creative system that cannot verify. And the hint or the database is also the key that makes this parallel possible. This is where tether matters. When you're prompting to an LLM or a text-to-image model just with the word Nongtong, you may get an average visual residue of every Nongtong room from this training data. It may be plausible, but it's belonging to no one in particular. And this is the place Corpus can be an anchor. Take an example. Writer Lu Da Feng recalls one evening in 1930s and written in his book. Sitting with friends in a long time courtyard, he heard the bamboo clapping after late night winter wonders and watch a girl in the building above lower, watch a girl in the building above lower a basket nugget from the stoking to make her purchase. The Wonton was thin-skinned, chicken-flavored, and served hot with egg strips and scallions. Stories from non-fictional writing like this add reliability and vividness to our corpus. writing like this add reliability and vividness to our corpus. So I extract passages like this from seven different sources from 1940s to 2020s and cut them into around 2,000 pieces. Each piece is tagged along five dimensions. Errors, life domain, space, emotion, and object is mentioned. Based on this, AI can create or remember vivid stories of Nong Thang from the perspective of daily practice and penetrate the layer of time. layer of time. And also, with this technical process, can we also create an experience? So that's where Anywhere Door is. Anywhere Door is an installation work in progress that's built on this idea. This concept is borrowed directly from that afternoon in the empty town version of Nong Town, where you can imagine the door inside from the fragment of the door outside. And you can choose a few objects from the panel and put it on the door. This is some sample of lists you can choose from. This is some sample of this you can choose from. Each object maps to tag and the tag maps back to the fragment in the corpus. A milk box and a spring couplet from different decades, an address place and a notice with different functions. The combination shifts the weight of the text. And then the system will retrieve several fragments based on this combination of text. And the language model connect and imagine them into our first person stories. Since the difference of the object chosen and storytelling is different each time, every visit produced a different narrations. And the physical structure is also emphasized in body experience. Audience will be curious when standing at the door and picking the object. Also, will be enlightened when hearing the story from the door and look through the peephole. And here is a video of the demo of this. And here is a video about the demo. when she was small probably 1990 something I could scrape it off but why the mailbox it's crooked now Thank you. 1935, said it made us a modern world. I was seven days, barely tall and I was touching a spell. And last part. These are the four different Nontown rooms you just saw in the videos. And this is the same act that's from the empty Nontown Alice and now can be formalized and made available to everyone. It is an invitation from a fictional permanent resident of Nongtang to approach the vanished Nongtang not through restoration or replica, but through the embodied act of imagining together with the machine and to think what it's like to be there. And there are some questions I'm still thinking about. First is about individual and collective. So here the map is the distribution of Nantang in 1947. That you can see is really the house for 70% of the Shanghai families. And I think it's collecting memory as a folk, a nebula made of countless fragments of individual memories. If the machine's imagination is layered on real sources, can it help us to across the folk, which cannot be taken apart to fully describe and came closer to the authenticity of one individual memory, rather than just casting the dust from the collective fork into a single shape. And second is about computation and embodiment. The physical interaction is part of the project, but it's also inside the remembering, the act of remembering itself. And in this work right now, the act of remembering or meditation is filled by the machine. If the purpose is to share the memory or even to educate visit, how can the visitor's own mind do more of the remembered effort themselves and with the involvement of AI. And that's all of my talk today. I hope maybe the experience can be shared with you in the short future. And that's a thank you. short future. And that's a thank you. Thank you for a great inspiring presentation. So have you already exhibited the installation and have you already got feedback? Not yet. I'm doing a demo online. Yeah, and building the physical part right now. So the next step to... Yes. Yeah. My question is, as I'm not an expert on the vanishing allies in Shanghai, so how do you distinguish between people who have a living there and know the allies, know the history and people who are not aware of the topic? Is that considered in an installation? Yes, because like I was mentioned, I do the on-site because like I was mentioned, I do the onsite investigation in Shanghai before I came to like US for work. But this idea came from like when already moved to California. And that's when I'm thinking of the idea, I always have questions with people around me. Like as someone who is not from Shanghai or from China, well, I'd be curious about this. Or is that really a matter that this kind of things is authentic to you or not? And I will say that maybe authentic, like for someone not from Shanghai, it's not really something matters. It's like everyone will love stories. And what I argue at the first is if you see from the current version of restored Nong Thang, like the building right now is only a slice of the past of Nong Thang. And it has demolished the other part that's the everyday life from this and I think like as Nong Thang goes through about the the history oh as the resident building the current structure or the current building is really lost the essence of the story so I think if for someone who is not interested in the history at all to hearing some interesting stories from the past that maybe you cannot imagine when you're not really living in this kind of building. Or being entertained with the stories is also my purpose. And so there's two sides. For someone who maybe lived there before, they can recall part of their memory from this process. And for the others, if they are interested in interact with this and hearing the stories that's maybe not happening today, is also very good things. Always when AI is involved and we have this balance between fiction and truth when it comes to memories and history, it is of course interpretation of memories, so to speak. Yeah. Yes. Great. So we have time for one last question from the audience. Any question? And thank you again for your presentation and all the presenters. Oh, here is one, sorry. Okay. Sorry. OK. Thank you for your great presentation. I'm stuck with the thought of, what was it? Remembering is an embodied act. And I'm very curious about your thoughts to it. Because I think, as I understood, the locations, they are gone. These doors are not there anymore. And I think, like, my hypothesis is that if you could display it still at these locations or have an experience about these memories at the locations, that would be for me personally the embodied act of remembering. So I'm curious about what your thoughts are. Did you think about that at all? Thanks for the question, and I think it will be great if there is an opportunity to really have this anywhere at the site. So as there is also the question just now about is this being authentic or not and have different meanings to different people. Yeah, I think for the more authentic of the memory or the experience, it would be really nice if we can instill these things back to the site. But only with this interactive experience, even outside the site, is the other way of creating stories or trying to make effort to picking things together and remembering things. And it's also, for me, it's also being meaningful in a part. Yeah. And if there is an opportunity, I would like to bring it back. Thank you. As we are running out of time and we need a break, again, thank you for your presentations and you are cordially invited to join us at 2.30 here again with the next art paper session. So thank you. Thank you.