Workshop: The Latent Spaces of Culture
About
Latent spaces are a foundational concept in contemporary artificial intelligence. They enable complex data—images, texts, sounds, and other digital objects—to be represented as high-dimensional structures in which relationships, patterns, and similarities can be modeled computationally. By translating intricate datasets into compressed mathematical representations, latent spaces allow AI systems not only to analyze information but also to generate new content; often, this content is itself cultural in nature. Yet the cultural and epistemological implications of latent spaces remain comparatively underexplored, particularly their role in shaping creativity, cultural memory, and representation. This three-day workshop, part of the France–Berkeley Fund project "The Latent Spaces of Culture," brings together researchers from UC Berkeley and institutions in France to examine these questions and explore latent space as both a technical architecture and a cultural form. (September 9, 2pm-4pm
