Blogs

 Berlin, April 10th 2024

Hello to all members of the Dance-tech.net family and enthusiasts of the digital dance realm,

As we look back on the remarkable journey that Dance-tech.net has undertaken since its inception in October 2007, it's thrilling to consider the path that lies ahead. Born from a vision to blend the realms of dance, technology, and digital culture, Dance-tech.net has grown into a vibrant community of artists, technologists, and scholars who share a passion for pushing the boundaries of what's possible in art and technology.

A Vision Expanded

Dance-tech.net has always been more than a platform; it's a living, breathing ecosystem that thrives on collaboration, innovation, and exploration. Our mission has led us to foster an environment where new ideas can flourish, where the intersection of movement and technology sparks creative fires.

Our Supportive Partners

As we embark on this next chapter, we're excited to announce the support of key partners who share our commitment to innovation and education in the realms of dance and technology. Lake Studios Berlin (lakestudiosberlin.com) continues to be a steadfast supporter, providing a space where artists and technologists can converge and create.

Radiona (radiona.org), with its dedication to open community and innovation, aligns with our vision of a collaborative future.

Furthermore, we're establishing special educational partnerships with leading organizations in the AI space: Hugging Face (huggingface.co) and Fal.ai (fal.ai/dashboard). These collaborations are set to enrich our community's understanding and application of AI in dance and performance, marking a significant leap forward in our collective exploration of digital technology.

Looking Forward

This next phase for Dance-tech.net is not just about growth; it's about deepening our engagement with the digital world and its potential to transform dance and artistic expression. We are standing at the threshold of a new era, ready to explore the possibilities of movement, embodiment, cognition and digital technology with a special focus on AI's role in redefining these domains.

An Invitation to Explore | Helps us  with new directions...

This journey would not be possible without you, our community. Your curiosity, passion, and creativity have been the driving force behind Dance-tech.net's evolution. As we move forward, I invite you to join us in this exciting new phase. Let's continue to challenge the boundaries of dance and technology, together. Let me know in the comments how do you feel about these new and how can we reacticate the community and these exciying filed of inquiry.

Warm regards,

Marlon Barrios Solano Curator | Dance-tech.net and Movimiento.org 

Play with this generative desgn and AI app here:

https://inpursuitoftolenghosts.vercel.app/

This is the unseeable space in which machine learning makes its meaning. Beyond that which we are incapable of visualizing is that which we are incapable of even understanding. — James Bridle

Latent space, in the realm of machine learning and artificial intelligence, refers to a high-dimensional abstract space where data’s intrinsic, hidden features are represented in a compressed form. This space is particularly significant in the context of generative models, such as Generative Adversarial Networks (GANs) and transformers, where it serves as the groundwork for the models to learn, encode, and manipulate the underlying patterns of the data they’re trained on. By navigating through this latent space, these models can generate new instances of data that, while reflective of the learned patterns, are distinct and original.

In “Duets in Latent Space,” embodies a live collaboration between the artist or user— situated before a laptop — and generative AI models that reside within this enigmatic latent space. Through various forms of input, whether they be movements, sounds, or digital interactions, the AI responds in kind, generating visual, auditory, or textual outputs that are played back in real time.

In “Duets in Latent Spaces,” I navigate different trajectories in the domain where machine learning finds its meaning inspired by James Bridle’s contemplation on locating these spaces beyond our capacity to visualize or understand.

“Duets in Latent Spaces” is conceptualized as a lecture-performance, installation and webpages, designed to bridge the realms of the tangible, the remembered and the speculative, facilitating presentations both in person and online.

This project unfolds through a series of vignettes combining stories, re-tellings, interfaces, software, movement scores and re-performances, each spanning 3 to 7 minutes, that illuminate trajectories of the generative potential of human-AI interaction, inviting the audience into a collaborative narrative that melds human intuition, storytelling with algorithmic playfulness.

Each segment is woven with personal stories, dances, histories about cybernetics, migration, post-coloniality and choreographic thinking. They venture into uncharted probability spaces, powered by bespoke applications I’ve developed to interface with advanced generative AI models and machine learning models for body and gesture recognition.

These interactions forge semantic and action landscapes that delve into the deep, unseen dimensions of data, cultural memory and language. By reversing the hegemonic narratives of generative AI and manipulating inputs, actions, and prompts, I navigate the AI models’ generative processes exploring their emancipatory potential, forging unique cognitive recombinations with evolving texts, images, and soundscapes set within the ethereal spaces of desire, affect, memory, longing and hybrid materiality.

Central to this project are several technologies: p5.js, enabling creative coding in the browser; Next.js, for rendering server-side React applications; machine learning models for hangs gesture and body movement recognition; humansLarge Language Models (LLMs), offering extensive capabilities for generating human-like text; and Generative Adversarial Networks (GANs), known for producing highly realistic images. These technologies underpin the performance, emphasizing REAL-TIME dynamic interactions that highlight the creative and epistemic challenges of generative AI.

Moreover, “Duets in Latent Spaces” extends beyond the performance realm into live, interactive applications, making the underlying technology directly accessible. The apps themselves, alongside their source code, are freely available, inviting further exploration and engagement from the audience and the broader community. This transparency aims to demystify the technology, encouraging a hands-on understanding and critical examination of AI’s role in shaping our perceptions of identity, creativity, and reality.

This project and each node is a mutable invitation to ponder the unseen forces that mold our digital and physical existence (as memory and imagination), offering a nuanced perspective on the intricate relationship between human cognition and artificial intelligence. Through this work, I challenge audiences to reflect on the invisible dimensions that influence our world, inhabiting the gap between the known and the unknowable latent space of entropic creativity.

It has been presented at:

-. Generative AI, Arts and Ethics, think-tank at Chateau de Fey, France March 5th of March 2024.

-. ACCAD Future Tech | The Ohio Stare University OSU Dance (Online) February 22nd 2024

-. Unfinished Fridays @ Berlin Lake Studios February 23rd 2024.

It can be adapted to be presented as installation-performances, installations, online apps and as line and online lecture-performance.

Nodes

Hyper-elements in Latent Space

• Model: Stable Diffusion XL
• Perlin Noise
• GitHub Repository with project description, code and live app

Extraordinary Alien

• Model: MediaPipeGesture Recognition
• GitHub Repository with project description, code and live app

Entropic Haiku

• Model: GPT-4
• GitHub Repository with project description, code and live app

In Pursuit of Stolen Ghosts

• Model: Stable Diffusion XL
• GitHub Repository with project description, code and live app

Bauhaus Time Traveller

• Model: Stable Diffusion XL
• GitHub Repository with project description, code and live app

Impossible Simplicity

• Model: MediaPipe Pose Landmark
• GitHub Repository with project description, code and live app

Lots of gratitude for the granted access of the most advanced realtime models from Fal.ai

For booking and/or information contact marlon@dance-tech.net

Reflecting on the culmination of our residency at Lake Studios Berlin in November 20th 2023, “Meta-fictions: A Fugue entangled with several odd kins and intelligences” represents a significant milestone in my journey exploring the realms of generativity, dance, and AI. This performance was not just the result of collaborative efforts but also a testament to the boundless possibilities that arise when different worlds of creativity and technology intersect.

Throughout the residency, my focus was on leveraging the capabilities of generative AI, to create an innovative choreographic tool also inrelation with the history of the use of rule systems and improvisational structures by dancers and actors. We iengaged in a dynamic dialogue with AI. We created a generative loop of continuous improvisation in collaborations with AIs. This process illustrated the true essence of generativity — an evolving, unpredictable artistic expression.

In “Meta-fictions,” we delved into the realm of self-aware storytelling. The performance was a conscious blend of fiction and reality, constantly evolving and reshaping itself based on the interactions between the dancers, AI, and the audience. This approach allowed us to explore the narrative in a way that questioned and played with the very nature of storytelling in the digital age. We collaborated and performed with fabulation engines.

A key aspect of the residency’s outcome was its rich tapestry of cultural influences due to our nationalities and complex histories: Venezuela, Mexico and Spain. From ancient indigenous chants to the modern rhythms of Spanish pop and American hits, the musical landscape of the performance was diverse and vibrant. The inclusion of Miguel Bose’s iconic performance in drag from Almodovar’s “Tacones Lejanos” brought an element of cinematic flair, merging music, film, and dance. The movements, inspired by the dramatic essence of bullfighting, connected our digital creation with deep-rooted cultural expressions. Longings, belongings and exiles were present.

In this project, AI took on multiple roles — as a repository of memories, a source of imagination, an evidence of cultural mashup and as collaging unique worlds. Durinf the process, AI-generated visuals of ethereal forms of jellyfish and algae, were not just artistic elements but symbolized the blending of the natural and digital realms, hight art conceptuality and fantasize about alien rituals. These OddKins represented the symbiosis of life forms. We watched Donna Haraway documentary and also sang karaoke. Working with LLMs is going meta all the time.

As I look back on “Meta-fictions,” I see it as an exploration in the field of performance art and installation. It’s a vivid example of how computation and AI happens is time and time happened in semantic realms. This residency at Lake Studios Berlin was an opportunity to navigate the uncharted (yet familiar) waters of generativity, AI, and dance improvisation. Stuff that changes in real-time.

Thanks to the artists collaborators in this residency Dakota Comin, Cesar Rene Perez, and Vera Rivas and the support of Lake Studios Berlin.

All images are from Maria Kousi and were taken in November 17th 2023.

Read experimental process documentation on Lake Studios Berlin blog