This paper presents VES AI – an innovative architecture for artificial intelligence based on the Viscous Emergent Spacetime (VES) theory and informational hydrodynamics. Unlike traditional neural networks, VES AI treats data as an informational fluid and the learning process as viscosity optimization. We describe the mathematical model (informational Navier-Stokes equations), the implementation of a VESLayer in PyTorch, and a cinematic generator built on top of it that creates video sequences from textual descriptions. Experiments confirm the approach’s efficiency (low computational cost, transparent decision-making) and its suitability for creative industries.
Mikheil Rusishvili (Wed,) studied this question.