<p><img src="https://www.ygzhang.com/images/edward/somenights-1.png" alt="Workshop Location"></p>
<p>A hands-on workshop diving into the landscape of text-to-image AIs with a special focus on Stable Diffusion, exploring new modes of content creation and examining our relationship with AI tools.</p>
<h2>Workshop Overview</h2>
<p>Over the past year, the unprecedented advancement in text-to-image artificial intelligence models has sparked widespread attention and discussion. This workshop examines both the excitement and concerns surrounding these tools, particularly focusing on open-source developments like Stable Diffusion.</p>
<h2>Key Topics Covered</h2>
<ul>
<li>Overview of text-to-image AI landscape</li>
<li>Deep dive into Stable Diffusion capabilities</li>
<li>Practical applications for artists and technologists</li>
<li>Ethical considerations and best practices</li>
<li>Hands-on experimentation with the tools</li>
</ul>
<h2>Technical Focus</h2>
<p>The workshop particularly emphasized:</p>
<ul>
<li>Stable Diffusion model architecture</li>
<li>Prompt engineering techniques</li>
<li>Custom model training possibilities</li>
<li>Integration with creative workflows</li>
<li>Resource optimization</li>
</ul>
<h2>Event Details</h2>
<p><strong>Date:</strong> November 3, 2022<br>
<strong>Time:</strong> 6:30–9pm ET<br>
<strong>Location:</strong> CultureHub NYC, 47 Great Jones Street<br>
<strong>Format:</strong> In-person workshop<br>
<strong>Cost:</strong> $25-60 Sliding Scale</p>
<h2>Teaching Artists</h2>
<h3>Tong Wu</h3>
<p>Creative technologist and multimedia artist focused on cyberculture and digital systems, exploring how individuals form relationships with their digital avatars in the AI era.</p>
<h3>Yuguang Zhang</h3>
<p>New media / AI artist working with interactive media, installation, and performance, investigating our connections with ubiquitous AI systems and the cultural shifts they bring.</p>
<p><em>Organized and hosted by CultureHub Inc., a global art &#x26; technology community.</em></p>


An introductory workshop exploring text-to-image AI models, focusing on Stable Diffusion's capabilities and implications for artists and creators

CultureHub

Text-to-Image with Stable Diffusion

<p><img src="/images/about/TongWu_YuguangZhang.webp" alt="Tong Wu and Yuguang Zhang"></p>
<p>We are <a href="https://tongwumedia.com/"><strong>Tong Wu</strong></a> &#x26; <a href="https://www.ygzhang.com/"><strong>Yuguang Zhang</strong></a>, a new media artist duo raised on the Internet, and the founders of <strong>Alternative Systems</strong>. We now co-exist with our digital doubles in Chrome browsers and Brooklyn, New York. Our joint artistic practice, which incorporates experiments with fine-tuned generative AI models, immersive installations, web-based interactions, 3D animations, community workshops, and performance, explores the dynamic relationships we could cultivate with autonomous intelligent systems, and the societal and cultural shifts that accompany these interactions.</p>
<h2>Our Approach</h2>
<p>As a duo, we share an interest in the notion of "alternatives," or what we call adjacent yet different scenarios from the present. Grounded in scientific principles, our artistic approach explores plausible alternative realities, especially those where humans and sentient machines may cohabit. We wonder if our view is only one story among many, and there could be other paths - accessible only to non-human intelligence - that reveal new visions of reality. Through our work, we invite the exploration of these alternative truths, and to reimagine our world in a broader, more inclusive light.</p>
<h2>Recognition &#x26; Awards</h2>
<p>Our work has been recognized through numerous prestigious awards and grants:</p>
<ul>
<li>National Endowment for the Arts (NEA) Grant, USA</li>
<li>Google AMI (Artist+Machine Intelligence) Research Award</li>
<li>The Lumen Prize Award (Still Image Category)</li>
<li>Digital Dozen Special Jury Awards, Columbia University</li>
<li>S+T+ARTS Project Award, EU</li>
<li>Re:Humanism AI Art Prize, Rome, Italy</li>
<li>MAX: Media Art Xploration Grant, New York, USA</li>
<li>Regional Emmy Award, USA</li>
</ul>
<h2>Exhibitions &#x26; Commissions</h2>
<p>Our art projects has been featured in significant venues and events such as:</p>
<ul>
<li>INDEX Biennale of Arts and Technology</li>
<li>City of Palo Alto Public Art Program</li>
<li>International Olympics Committee (IOC) Young Leaders Programme</li>
<li>MAXXI - National Museum of XXI Century Arts, Rome</li>
<li>Le Centquatre, Paris (Némo Biennial of Digital Arts)</li>
<li>New York Live Arts</li>
<li>Museum of Science, Boston</li>
<li>CURRENTS New Media Festival, Santa Fe</li>
</ul>
<h2>Academic &#x26; Research Positions</h2>
<p>We maintain active roles in academia and research:</p>
<ul>
<li>Machine Learning Research Fellowship at NYU ITP</li>
<li>Adjunct Professor, NYU Tisch School of Arts (Teaching Generative AI Art)</li>
<li>Research Fellow, Interactive Media Arts LowRes, NYU</li>
<li>Artist Residencies at MAX: Media Art Xploration, Danspace Project, and SloMoCo</li>
</ul>
<h2>Speaking &#x26; Workshops</h2>
<p>We regularly share our expertise through talks and workshops at prestigious institutions such as:</p>
<ul>
<li>Salzburg Media Academy</li>
<li>Sundance Artist Accelerator Program</li>
<li>CUNY Graduate Center Art Science Connect</li>
<li>School for Poetic Computation (SFPC)</li>
<li>Yale University CCAM</li>
<li>School of Visual Arts (SVA)</li>
<li>BlackStar Film Festival</li>
<li>International Documentary Film Festival Amsterdam (IDFA)</li>
<li>CultureHub</li>
</ul>
<h2>The Artists</h2>
<p><strong>Tong Wu</strong> holds a BA degree from the University of Maryland, College Park, and a Master's degree from the Interactive Telecommunications Program (ITP) at New York University. She is an AI research engineer, whose research work uses design, code and narrative to explore the possibilities and boundaries of intelligent systems experience design.</p>
<p><strong>Yuguang Zhang</strong> is a machine learning technologist and a member of interdisciplinary collective NUUM which creates research-based performance with movement, sound, text, visual art and coded rules of interaction. He is also an adjunct professor of the Interactive Telecommunications Program (ITP) at New York University's Tisch School of Arts, where he teaches courses about generative AI for creative media production.</p>
<h2>Press &#x26; Media</h2>
<p>Our work has been featured in various publications including:</p>
<ul>
<li>Neural Magazine</li>
<li>Variety Magazine</li>
<li>The Paper.cn</li>
<li>Al-Tiba9 Contemporary Art</li>
<li>ARTRON</li>
<li>ReadCopy.co</li>
</ul>


About

<p><img src="/images/projects/alternative-family-stories/Album.png" alt="Alternative Family Stories Album"></p>
<p>Alternative Family Stories explores the boundaries between authentic and artificial memories through AI-generated family photographs. By fine-tuning text-to-image models with personal photos, we create speculative family albums where AI's inherent inconsistencies become a feature that challenges our understanding of identity and memory.</p>
<h2>Concept &#x26; Story</h2>
<p>In an era where artificial intelligence increasingly shapes our visual world, the line between authentic and generated imagery becomes increasingly blurred. This project embraces the imperfections of AI generation, turning them into a narrative device that questions how we document and remember human experiences.</p>
<p>The project presents AI-generated portraits as vintage family photos in various physical forms - from traditional photo albums to framed portraits on old television sets. These artifacts create a bridge between traditional methods of preserving memories and the new frontier of AI-generated content.</p>
<p><img src="/images/projects/alternative-family-stories/album_desk.jpg" alt="Family Portrait Series"></p>
<h2>Exhibition Components</h2>
<h3>Physical Installation</h3>
<p>The centerpiece of the exhibition features carefully curated photo albums and framed portraits, each telling a different speculative family story. These physical artifacts are designed to be indistinguishable from genuine vintage photographs, creating an uncanny valley effect that prompts viewers to question their assumptions about authenticity.</p>
<p><img src="/images/projects/alternative-family-stories/portrait_tv.png" alt="Installation View"></p>


An experimental project that uses AI-generated portraits to create fictional family narratives

Speculative family albums exploring AI-generated personal histories

Alternative Family Stories

<p><img src="https://www.ygzhang.com/images/cloud/Inside_paloalto_version.png" alt="BCM Installation View"></p>
<p>The Bureau of Cloud Management (BCM) is a government agency of the alternate reality that oversees the entry of clouds into its designated monitoring area. It is a simulated environment where we recreate our shared memory of clouds with cloud technology.</p>
<h2>Concept &#x26; Story</h2>
<p>In an era when the "cloud" is interpreted as the core concept of global-scaled network computing systems, it represents advancement - faster speed, broader access, higher efficiency, and greater connection. While the world enjoys the adrenaline brought by technical revolution, feelings of fatigue and languishing are also widely experienced as we try to catch up with technology adaptation and overloaded information.</p>
<p>Are we closer to or further away from being free? Driven by this thought, we trace back to the original cloud - and its characteristics associated with our primitive thinking of and connection with the world.</p>
<p>Perhaps every human being has had a moment of looking up at the clouds - those afternoons of laying on the lawn and simply observing the fluffy, cotton-like clouds slowly floating by, constantly changing shapes. These moments draw the initial picture of the dreams of every one of us. Building upon countless gazes and primitive human imagination across time and space, clouds constitute the broadest interconnection of the world while embodying an ultimate form of unrestricted freedom.</p>
<p><img src="https://www.ygzhang.com/images/cloud/BureauOfCloudManagement_01.jpg" alt="BCM Inner Scene"></p>
<h2>Technical Implementation</h2>
<p>The project utilizes various technologies and tools:</p>
<h3>Generation &#x26; Training</h3>
<ul>
<li>UMAP and Word2Vec (2021), Arctic / Cohere / Gemini (2023-) for embeddings</li>
<li>Llava and GPT / Gemini families for text generation &#x26; image-to-text conversion</li>
<li>CLIP Guided Diffusion (2021), Stable Diffusion (2022-), Flux (2024-) for image generation</li>
</ul>
<p><img src="https://www.ygzhang.com/images/cloud/clouds.png" alt="Cloud Generation Process"></p>
<h3>Interactive Development</h3>
<ul>
<li>Three.js for web version</li>
<li>Unreal Engine 5 for Virtual Reality version</li>
<li>Cinema 4D for scene building</li>
</ul>
<p><img src="https://www.ygzhang.com/images/cloud/ue5.gif" alt="VR Experience"></p>
<h2>Recent Exhibitions</h2>
<p><img src="https://www.ygzhang.com/images/cloud/BCM_iidrr1.jpg" alt="Exhibition at :iidrr Gallery">
<img src="https://www.ygzhang.com/images/cloud/BCM_iidrr2.jpg" alt="Exhibition at :iidrr Gallery"></p>
<ul>
<li>We May Be in <em>This</em> Together, Duo Exhibition, :iidrr Gallery, New York, USA (Nov 2023)</li>
<li>Code:ART, City of Palo Alto, USA (Oct 2023)</li>
<li>Alpha, :iidrr Gallery, New York, USA (Dec 2022)</li>
<li>INDEX Biennial of Art and Technology, Braga, Portugal (May 2022)</li>
<li>Machine Learning for Creativity and Design, NeurIPS 2021</li>
</ul>
<p><img src="https://www.ygzhang.com/images/cloud/BCM_Palo_Alto.jpg" alt="Installation at Palo Alto">
<img src="https://www.ygzhang.com/images/cloud/audience1.JPG" alt="Installation at Palo Alto"></p>
<h2>Project Components</h2>
<h3>Web Portal</h3>
<p>The online version of the project is a web-based experience that presents the monitoring interface of BCM. Participants will be given the identity as a BCM worker and get to record and archive the passing-by clouds. It is one of the four online artworks selected in the INDEX 2022 Open Call and part of the INDEX Biennale 2022 in Braga, Portugal.</p>
<p>We processed the collected folklores and local cultural reports about Braga using GPT-J, an open-source text-generating ML model. Keywords with visual reference were extracted as the text prompts to be fed into a tuned CLIP-Guided Diffusion model to generate various 2D cloud shapes. The cloud shape images were later processed by a custom computer vision pipeline and turned into textures for rendering the slowly floating clouds on the website using Three.js.</p>
<p>The portal is accessible at <a href="https://clouds.management/">https://CLOUDS.MANAGEMENT/</a></p>
<p><img src="/images/projects/bcm/web_portal.png" alt="BCM Web Interface">
<img src="https://www.ygzhang.com/images/cloud/BCM_InnerLook_TongWuYuguangZhang.jpg" alt="BCM Web Interface">
<img src="https://www.ygzhang.com/images/cloud/BCM%20Inner%20Scene%20Desktop.jpg" alt="BCM Web Interface"></p>
<h3>VR Experience</h3>
<p>The VR prototype uses real cloud trajectory data observed in Brooklyn as the path for finding "cloud prompts" in UMAP embeddings of Word2Vec. Generated clouds are converted into 3D assets for the immersive Bureau environment.</p>
<p><img src="https://www.ygzhang.com/images/cloud/map.png" alt="BCM Interface"></p>
<h3>Community Engagement</h3>
<p>The project maintains an <a href="https://www.instagram.com/manage_clouds/">official Instagram account</a> for remote cloud reports from observers worldwide, creating a growing community of cloud observers.</p>
<p><img src="https://www.ygzhang.com/images/cloud/audience2.jpg" alt="Audience Interaction"></p>


An imaginary government agency that oversees the entry of clouds into its designated monitoring area, exploring the intersection of natural freedom and systematic control

A simulated environment where clouds meet bureaucracy

Bureau of Cloud Management

<p><img src="/images/teaching/creating-beyond-words/cover.png" alt="Workshop Session"></p>
<p>A comprehensive three-session workshop series introducing design students to the world of generative AI, exploring practical applications and ethical considerations in creative practice.</p>
<h2>Workshop Structure</h2>
<h3>Session 1: Foundations</h3>
<ul>
<li>Introduction to Generative AI</li>
<li>History of Image Synthesis and Generation</li>
<li>Evaluating Generative AI Models</li>
<li>Hands-on Text-to-Image Generation Practice</li>
</ul>
<h3>Session 2: Advanced Applications</h3>
<ul>
<li>Value Assessment of AI Tools in Design</li>
<li>Introduction to Dreambooth</li>
<li>Advanced Text-to-Image Techniques</li>
<li>Practical Exercises</li>
</ul>
<h3>Session 3: Ethics and Future Directions</h3>
<ul>
<li>Reflection on AI Tool Usage</li>
<li>Text-to-Video Technologies</li>
<li>Ethical Boundaries in AI Generation</li>
<li>Discussion of Best Practices</li>
</ul>
<h2>Technical Coverage</h2>
<p>The workshop series explored:</p>
<ul>
<li>Text-to-image generation techniques</li>
<li>Prompt writing strategies</li>
<li>2D &#x26; 3D Animation with AI</li>
<li>Video interpolation methods</li>
<li>Video-to-video transformation</li>
</ul>
<h2>Learning Outcomes</h2>
<p>Students gained practical knowledge in:</p>
<ul>
<li>Evaluating AI tools for design work</li>
<li>Creating effective prompts</li>
<li>Understanding ethical considerations</li>
<li>Integrating AI into creative workflows</li>
<li>Critical thinking about AI application</li>
</ul>
<h2>Teaching Team</h2>
<h3>Tong Wu</h3>
<p>Creative technologist exploring the intersection of AI and design</p>
<h3>Yuguang Zhang</h3>
<p>New media artist specializing in AI and interactive installations</p>
<p><em>Organized for second-year Interaction Design students</em></p>


A three-session workshop series exploring generative AI tools for design, focusing on text-to-image and text-to-video technologies

SVA MFA Design

Creating Beyond Words

<p><img src="https://images.squarespace-cdn.com/content/v1/57e4a5caf5e231b14d767846/1623973018233-D1DF6H3ZMVVOBP3YJ0HM/Exhibition_wide_02.jpg" alt="Exhibition View"></p>
<p>Daily Dividuals is an artistic research project investigating our relationship with our digital doubles, documenting observations about the expanding philosophical brokerage between the physical world, the virtual sphere, and online avatars.</p>
<h2>Concept &#x26; Research</h2>
<p>In etymology, "Individual" traces back to "not to be divided" in Latin. As technological advancement enables us to see beyond the naked eye, philosopher Gilles Deleuze suggests this term no longer conveys society's smallest reducible unit. In our machine-dominated society, individuals are metamorphosing into data masses while our online avatars - our 'Dividuals' - actively engage in the fragmented virtual world.</p>
<p><img src="https://images.squarespace-cdn.com/content/v1/57e4a5caf5e231b14d767846/1623973064203-KYYS89IX2RAEXPSY4YTB/DailyDividuals_I_am_Your_Fan_installation_02.jpg" alt="Sculpture Installation"></p>
<h2>Technical Implementation</h2>
<p>The project experiments with multiple media forms:</p>
<ul>
<li>3D animated videos</li>
<li>Augmented Reality (AR) installations</li>
<li>Digital sculpture installations</li>
<li>Virtual exhibition spaces</li>
<li>Digital paintings</li>
</ul>
<p><img src="https://images.squarespace-cdn.com/content/v1/57e4a5caf5e231b14d767846/1623968117399-D2HWBLNHJFKSNBMGJQ9N/DailyDividuals+Online+Space+Poster.jpg" alt="Virtual Exhibition"></p>
<h2>Exhibition Components</h2>
<h3>Virtual Exhibition Platform</h3>
<p>A custom-designed virtual exhibition space created for the solo exhibition at LATITUDE Gallery, New York, enabling remote access and digital interaction with the works.</p>
<h3>Physical Installations</h3>
<p>Multiple installation pieces including:</p>
<ul>
<li>"I am your FAN" - 3D printed sculpture series</li>
<li>AR installations exploring human-machine perception</li>
<li>Digital animation sequences</li>
</ul>
<p><img src="https://images.squarespace-cdn.com/content/v1/57e4a5caf5e231b14d767846/1623965527452-OQV62VBADRWS5XQGJ91G/DailyDividuals_HumanSample5237_AR_installation.jpg" alt="AR Installation"></p>
<h2>Exhibition History</h2>
<ul>
<li>LATITUDE Gallery, New York, USA (Solo Exhibition, 2021)</li>
<li>Various virtual exhibitions and online platforms</li>
<li>Multiple AR installations in public spaces</li>
</ul>
<p><img src="https://images.squarespace-cdn.com/content/v1/57e4a5caf5e231b14d767846/1625441850909-5KFPK90N7VCG7DVD3DD8/DailyDividuals_ConceptualPoster_011.png" alt="Conceptual Visual"></p>


An ongoing exploration of our relationship with digital avatars, examining the expanding philosophical brokerage between physical and virtual existence

An artistic research on digital doubles

Daily Dividuals

<p>Edward is a series of short animation experiments about a speculative world interpreted via the eyes of Edward Hopper. All animations are created using a custom-built image synthesis pipeline that applies multiple layers of interpolation and blending techniques on top of images generated via Stable Diffusion.</p>
<h2>Series Components</h2>
<h3>Edward - I : Some Nights</h3>
<p>A visual exploration of nocturnal scenes reminiscent of Hopper's iconic nighttime paintings.</p>
<p><img src="https://www.ygzhang.com/images/edward/somenights-1.png" alt="Some Nights Scene 1">
<img src="https://www.ygzhang.com/images/edward/somenights-2.png" alt="Some Nights Scene 2">
<img src="https://www.ygzhang.com/images/edward/somenights-3.png" alt="Some Nights Scene 3">
<img src="https://www.ygzhang.com/images/edward/somenights-4.png" alt="Some Nights Scene 4"></p>
<h3>Edward - II : Bob</h3>
<p>A character study exploring solitude and urban spaces through generated imagery.</p>
<p><img src="https://www.ygzhang.com/images/edward/Bob-2.png" alt="Bob Scene 1">
<img src="https://www.ygzhang.com/images/edward/Bob-6.png" alt="Bob Scene 2">
<img src="https://www.ygzhang.com/images/edward/Bob-5.png" alt="Bob Scene 3">
<img src="https://www.ygzhang.com/images/edward/Bob-3.png" alt="Bob Scene 4"></p>
<h3>Edward - III : Roll</h3>
<p>An experimental sequence investigating movement and transition in Hopper-inspired spaces.</p>
<p><img src="https://www.ygzhang.com/images/edward/Roll-1.png" alt="Roll Scene 1">
<img src="https://www.ygzhang.com/images/edward/Roll-2.png" alt="Roll Scene 2">
<img src="https://www.ygzhang.com/images/edward/Roll-3.png" alt="Roll Scene 3">
<img src="https://www.ygzhang.com/images/edward/Roll-4.png" alt="Roll Scene 4"></p>
<h2>Technical Implementation</h2>
<p>The project employs a sophisticated technical pipeline:</p>
<ul>
<li>Custom-built image synthesis system</li>
<li>Multiple layers of interpolation and blending techniques</li>
<li>Stable Diffusion as the base image generation model</li>
<li>Animation techniques for smooth transitions between generated states</li>
</ul>


A series of short animation experiments created using AI-powered image synthesis, interpreting reality through Edward Hopper's artistic lens

A speculative world through Edward Hopper's eyes

Edward

<p><img src="https://images.squarespace-cdn.com/content/v1/57e4a5caf5e231b14d767846/1591728952778-GSBWX1S3VNXMS88KSUFJ/LiveLab+Landing+Page" alt="LiveLab Interface"></p>
<p>LiveLab is an open-sourced, browser-based media routing software designed for online collaborative performance, developed at CultureHub Inc., a global art &#x26; technology community founded by SeoulArts &#x26; La MaMa.</p>
<h2>Project Overview</h2>
<p>As Creative Technologist at CultureHub Inc., I contributed to developing this innovative platform that democratizes access to telecommunication tools for creators worldwide. The software expands current offerings by allowing users to customize media routing in ways that best suit their creative needs.</p>
<p><img src="https://images.squarespace-cdn.com/content/v1/57e4a5caf5e231b14d767846/1616193079064-2ETHRKYWV8F61JNM8YJX/CultureHub+x+La+Mama+logo" alt="CultureHub Partnership"></p>
<h2>Technical Implementation</h2>
<p>Built with modern web technologies:</p>
<ul>
<li>React.js frontend framework</li>
<li>WebRTC for real-time communication</li>
<li>HTML5/CSS3 for interface design</li>
<li>Git for version control</li>
<li>Webpack for bundling</li>
<li>Figma for UI/UX design</li>
</ul>
<h2>Core Features</h2>
<h3>Media Management</h3>
<ul>
<li>Multiple camera and audio stream support</li>
<li>Dynamic feed addition/removal</li>
<li>Customizable video resolution</li>
<li>Advanced audio processing options</li>
</ul>
<h3>Performance Tools</h3>
<ul>
<li>Audio mixer with individual feed control</li>
<li>Video switcher for multi-camera setups</li>
<li>Pop-out windows for external hardware integration</li>
<li>Built-in chat system</li>
</ul>
<p><img src="https://images.squarespace-cdn.com/content/v1/57e4a5caf5e231b14d767846/1616194238208-28CHXH01GNIB4EM75AYI/vital+signs.jpeg" alt="Performance Example"></p>
<h2>Project Links</h2>
<ul>
<li><a href="https://livelab.app/">Live Platform</a></li>
<li><a href="https://github.com/CultureHub/LiveLab">GitHub Repository</a></li>
<li><a href="https://www.culturehub.org/livelab">Project Documentation</a></li>
</ul>
<h2>Team Credits</h2>
<ul>
<li>Lead Software Developer: Olivia Jack</li>
<li>Creative Technologist: Tong Wu</li>
<li>Additional Contributors: Jesse Ricke, Damián Silvani</li>
</ul>
<p><em>Supported by CultureHub Inc. and La Mama Experimental Theatre Club</em></p>


A collaborative performance software enabling artists to meet, create, and produce multi-location performances from anywhere in the world

Open-source browser-based media routing platform

LiveLab

<p><img src="/images/projects/nephophile/0.jpg" alt="Nephophile Installation View"></p>
<p><em>"Nephophile" are machines that observe clouds.</em></p>
<p><em>They are protagonists from Yuguang Zhang and Tong Wu's world-building project "<a href="/projects/bureau-of-cloud-management">Bureau of Cloud Management</a>" — an imagined alternate reality where human and AI cloud observers (codename "Nephophile") are both tasked with observing and documenting clouds under systematized workflow.</em></p>
<h2>Cloud Observing Machines</h2>
<p>In the wild field of Maajaam, Estonia, three Nephophile devices, each consisting of a camera, a phone booth and an AI agent connected between them, are scattered across the lawn. They are positioned to study clouds differently—one sees through trees, one watches the pond, and one views the open sky.</p>
<p><img src="/images/projects/nephophile/1.jpg" alt="Nephophile Device in Field"></p>
<p><img src="/images/projects/nephophile/3.jpg" alt="Nephophile Device Close-up"></p>
<p><img src="/images/projects/nephophile/2.jpg" alt="Nephophile Camera Detail"></p>
<p>Each day, the three devices capture photos of clouds in their respective environments. Inside each device, an AI agent composed of a VLM and an LLM with knowledge of "cloud" suppressed, is tasked to digest such captures and develop new understanding of clouds. The captured scenes over time gradually accumulate and become their primary sources for learning what "clouds" could be.</p>
<p>The paired phone booth to each camera is the outlet for visitors to catch a snippet of this ongoing observation. Visitors can lift telephones to hear how situated experiences gradually shape divergent yet intimate definitions of "clouds."</p>
<p><img src="/images/projects/nephophile/5.jpg" alt="Nephophile Phone Booth"></p>
<h2>Technical Implementation</h2>
<p>The project utilizes various technologies and tools:</p>
<ul>
<li>ESP32 microcontrollers</li>
<li>Custom Electronics</li>
<li>Stainless Steel construction</li>
<li>3D Printed components</li>
<li>No-Dial Wall Phones</li>
<li>Custom AI agents (VLM + LLM)</li>
</ul>
<h2>Exhibition View @ Wild Bits, Maajaam, Ojaveere, Estonia</h2>
<p><img src="/images/projects/nephophile/6.jpg" alt="Exhibition View 1"></p>
<p><img src="/images/projects/nephophile/6-1.jpg" alt="Exhibition View 2"></p>
<p><img src="/images/projects/nephophile/6-2.jpg" alt="Exhibition View 3"></p>
<p><img src="/images/projects/nephophile/7.jpg" alt="Exhibition View 4"></p>
<p><em>All photos by Ya Chuan Chen (J.)</em></p>
<h2>Awards &#x26; Exhibitions</h2>
<ul>
<li><strong>Wild Bits Art &#x26; Tech Grant</strong>, Maajaam, Ojaveere, Estonia</li>
<li><strong>Wild Bits 2025</strong>, Maajaam, Ojaveere, Estonia</li>
</ul>


Nephophile are machines that observe clouds - protagonists from the world-building project 'Bureau of Cloud Management' where human and AI cloud observers are tasked with observing and documenting clouds under systematized workflow

Cloud observing machines in the field

Nephophile

<p><img src="https://www.ygzhang.com/images/non-human/1.jpg" alt="Installation View"></p>
<p>(Non-)Human is an art installation that conjures the hidden human-ness in objects and imagines a speculative world where a human exists in non-human forms.</p>
<h2>Awards &#x26; Recognition</h2>
<ul>
<li>Re:Humanism AI Art Prize, 2021</li>
<li>S+T+ARTS S2S Award Winning Work, 2021</li>
<li>CVPR Art Gallery 2021 Shortlist</li>
</ul>
<h2>Concept &#x26; Research</h2>
<p>We live a life surrounded by objects we build to serve us. While we tend to think of ourselves as different - possessing spirituality, reason, and intelligence - objects keep our traces, carry our memories, and shape us both physically and behaviorally. The project explores this deeper connection between humans and objects that often goes unnoticed in daily life.</p>
<p><img src="https://www.ygzhang.com/images/non-human/spectrum.jpg" alt="Concept Visualization"></p>
<p>The research spans three domains:</p>
<ul>
<li>Physics and cosmology regarding the origin of life</li>
<li>Philosophical notions about consciousness</li>
<li>Religious roots including Buddhism and Shinto</li>
</ul>
<h2>Technical Implementation</h2>
<p>The installation utilizes various technologies and components:</p>
<ul>
<li>Arduino and stepper motors for mechanical movement</li>
<li>Custom GAN trained on 300+ hours of sleep recordings</li>
<li>PoseNet and RealSense for movement tracking</li>
<li>MongoDB for movement data storage</li>
<li>Custom bed frame with eight motor joints</li>
</ul>
<p><img src="https://www.ygzhang.com/images/non-human/bedframe2.gif" alt="Technical Process"></p>
<h2>Movement Studies</h2>
<p>The project began with studying human sleep patterns, identifying three major movement categories:</p>
<ul>
<li>Side Curl</li>
<li>Laying Flat</li>
<li>Roll-over</li>
</ul>
<p><img src="https://www.ygzhang.com/images/non-human/movement1.gif" alt="Movement Study"></p>
<h2>Exhibition History</h2>
<ul>
<li>Némo Biennial of Digital Arts, Oct 2021</li>
<li>Beijing Times Art Museum, Aug 2021</li>
<li>Maxxi - The National Museum of XXI Century Arts, May 2021</li>
<li>NeurIPS Creativity Workshop, Dec 2020</li>
<li>The SHIM/Masters Show, Jul 2020</li>
</ul>
<p><img src="https://www.ygzhang.com/images/non-human//19.jpg" alt="Exhibition View"></p>
<p><em>Photo by Sebastiano Luciano; courtesy Re:Humanism.</em></p>


A speculative world where human exists in non-human forms, examining the blurred boundaries between human and object through an intelligent bedsheet installation

Kinetic AI installation

(Non-)Human

<p><img src="/images/projects/ioc2024/cover.jpg" alt="thumbnail"></p>
<h2>Project Overview</h2>
<p>Quanzhou Move Paris Dream is a commissioned AI-generated video work celebrating Olympic Day 2024, created for IOC Young Leader Shiling Lin's GreenNwild initiative. The video explores the intersection of breaking culture - newly added to the Paris 2024 Olympics - with traditional Chinese cultural elements, demonstrating how sport transcends cultural boundaries.</p>
<h2>Concept &#x26; Vision</h2>
<p>The project aims to convey breaking's unique fusion of art, sport and culture in a way that breaks boundaries, while illustrating that through sport, we all speak the same language. The narrative bridges two distinct cultural spaces: Quanzhou, a city rich in traditional Chinese culture, and Paris, the host of the 2024 Olympics where breaking will make its debut as an Olympic sport.</p>
<p><img src="/images/projects/ioc2024/ioc2024-screenshot-1.png" alt="Video Screenshots"></p>
<h2>Technical Implementation</h2>
<p>The video utilizes advanced AI techniques to create seamless transitions between scenes, which fluidly shift between China and Paris. The technical approach includes:</p>
<ul>
<li>Integration of traditional Chinese ink wash paintings</li>
<li>Implementation of impressionistic art styles</li>
<li>AI-powered scene transitions that blur the boundaries of reality</li>
<li>Seamless visual narrative connecting Eastern and Western cultural elements</li>
</ul>
<h2>Visual Language</h2>
<p>The work employs various artistic styles to create a cohesive visual narrative:</p>
<ul>
<li>Traditional Chinese ink wash painting aesthetics</li>
<li>French impressionism influences</li>
<li>Contemporary breaking culture visual elements</li>
<li>Dynamic transitions that transcend geographical boundaries</li>
</ul>
<p><img src="/images/projects/ioc2024/cover_2.png" alt="thumbnail"></p>
<h2>Cultural Impact</h2>
<p>This project represents a significant intersection of:</p>
<ul>
<li>Traditional cultural heritage and contemporary urban sports</li>
<li>Olympic values and local cultural expression</li>
<li>Technological innovation in artistic presentation</li>
<li>Cross-cultural dialogue through sport and art</li>
</ul>
<p>The video serves as a bridge between traditional Chinese culture and modern Olympic sports, demonstrating how AI technology can be used to create meaningful cultural connections while promoting the universal language of sport.</p>
<h2>Exhibition</h2>
<p>The video was featured as part of the Olympic Day 2024 celebrations, showcasing how sport can break boundaries between cultures and create spaces where everyone can enjoy sport, friendship, and peace.</p>


A commissioned video work celebrating Olympic Day 2024, using AI to create seamless transitions between Quanzhou and Paris, blending traditional Chinese art with contemporary breaking culture

Video Art

Quanzhou Move Paris Dream

<p><img src="/images/teaching/t2i/cover.png" alt="Course Banner"></p>
<p>A semester-long exploration of Text-to-Image &#x26; Text-to-Video AI technologies at New York University's Interactive Telecommunications Program (ITP) since 2023, covering everything from fundamental concepts to advanced applications in animation and interactive storytelling.</p>
<h2>Technical Focus</h2>
<p>The course covers:</p>
<ul>
<li>Open source text-to-image models like Stable Diffusion / Flux</li>
<li>Custom model training</li>
<li>Advanced prompt engineering</li>
<li>Animation workflows</li>
<li>Interactive implementations</li>
</ul>


A comprehensive course exploring text-to-image AI models, from basic generation to advanced customization and animation