Principia Artificialis Open mathematical research program onthe theory of AI (github.com)

Principia Artificialis has launched an innovative open research program dedicated to the mathematical underpinnings of artificial intelligence (AI). By harnessing advanced theoretical physics concepts—such as information geometry, topology, and thermodynamics—the initiative aims to frame machine intelligence not merely as software, but as a physical theory. This shift is significant for the AI/ML community as it opens new pathways to rigorously measure and understand intelligence through quantifiable metrics akin to physical quantities, theoretically allowing for the exploration of scaling laws and phase transitions. The project encourages collaboration and contribution from researchers, with initial hypotheses that delve into topics like the "Curvature of Reasoning," which examines reasoning trajectories on information manifolds, and the "Entropy-Production based depth of inference," likening complex thought processes to physical phenomena. As all contributions remain in draft status, there is a clear emphasis on maintaining a scientific rigor around proposed ideas, compelling the AI/ML community to explore and validate these cutting-edge concepts. Overall, Principia Artificialis aims to propel a deeper understanding of AI’s cognitive capabilities through a mathematical lens, fostering a collaborative environment for innovative research and discourse.