Abstract:
Current computational models, rooted in the axiomatic framework of the Turing machine, treat information as a one-dimensional, context-agnostic sequence of symbols. This paradigm, while foundational to the digital age, is increasingly strained by the complex, high-dimensional, and context-rich data that fuel modern artificial intelligence. This paper introduces the Spatiotemporal Information Framework (STIF), a novel theoretical model that re-axiomatizes data as a fundamentally four-dimensional entity. Within this framework, a elemental unit of information, termed an “informon,” is formally defined as a rank-4 tensor residing within a dynamic computational manifold. Its four dimensions correspond to Locus (Iλ, the value), the Contextual Manifold (Iκ, relational structure), the Temporal Vector (Iτ, causal history), and the Metastatic Field (Iμ, potentiality and uncertainty). This re-conceptualization recasts algorithms not as sequential instructions but as the search for geodesics within this computational spacetime, where the “cost” of an operation is intrinsically linked to the local “curvature” induced by information density and complexity. STIF offers a native architecture for truly context-aware AI, provides a theoretical basis for emergent properties such as common-sense reasoning and robust memory, and proposes a profound unification of computational theory with foundational principles from general relativity and quantum mechanics. It suggests that information and computation are not abstract logical processes but are, in fact, fundamentally physical phenomena governed by geometric principles.
Yıldırım, E. (2025). The Spatiotemporal Information Framework (STIF): A Relativistic Model of Computation. Zenodo. https://doi.org/10.5281/zenodo.17044982
Leave a Reply