At NVIDIA GTC 2024, Refik Anadol presented an unprecedented blend of nature and imagination through human-AI collaboration, captivating the audience and setting the stage for Jensen Huang's much-anticipated keynote. After a stunning premiere at the World Economic Forum in Davos, immersive artworks based on Refik Anadol Studio’s Large Nature Model (LNM) came to the US for the first time at the NVIDIA GTC 2024. The Large Nature Model transforms more than 100 million images of the Earth's diverse flora, fauna, and fungi into breathtaking visuals reminiscent of an Amazonian rainforest. "In the realm of AI, human intelligence and reasoning have dominated the conversation. Yet, I believe nature itself offers a fresh, invaluable perspective that we're just beginning to explore," Anadol remarked, hinting at the groundbreaking potential of his project. Large Nature Model: A Living Archive was situated prominently on the main concourse of the San Jose Convention Center, where the global AI event took place, from March 18-21.
Anadol's journey in AI and art has been marked by significant collaborations. Reflecting on the support that fueled his endeavors, Anadol highlighted, "The last eight years have been transformative, with significant collaborations starting from Gamescom and extending to NVIDIA’s robust ecosystem, including DGX stations and A6000 graphics units. These tools have been instrumental in managing and interpreting big data locally." Moreover, NVIDIA's Omniverse and Picasso software ecosystems played a critical role in his ability to curate and refine half a billion open-source nature images from prestigious institutions like the Smithsonian, Natural History Museum, and National Geographic. In addition to the open archives of their data partners, Anadol’s team is venturing to 16 unique rainforest locations, deploying data collection technologies ranging from lidar to photogrammetry, and capturing ambisonic audio and high-resolution visuals of these diverse ecosystems.
Anadol's installation at GTC 2024 took attendees on an unparalleled multi-sensory journey, engaging not just sight and sound but also the sense of sound and smell. The artworks' realistic and immersive soundscapes were generated and curated by using cutting-edge pre-trained audio neural networks and a custom audio latent diffusion model trained on a broad spectrum of complex nature sounds. By dreaming up the vibrant flora of the Amazon through AI, visitors were also invited to "smell the AI dreams," an innovative experience that underscores Anadol's belief in the profound impact of scent on memory and experience. At the heart of this olfactory innovation are two advanced AI systems working in tandem: one AI, trained on about half a million scent molecules, and another, the Large Nature Model, which holds the world's biggest collective dataset on the flora, fauna, and fungi. As the artist introduced the two AI systems to each other, merging the database of scent molecules with the biodiversity knowledge from the LNM, the Scent AI became able to generate formulas that recreated the almost realistic scents of nature.
Refik Anadol Studio's foray into interactive AI research brings to life generative realities that combine real-time rendering with sensory inputs. This innovative project envisions fully interactive 3-D worlds crafted by the collaboration between AI and the artist. It invites the audience to explore millions of raw datasets from the Large Nature Model.