The Universe as a Self-Learning Relational Network: Analogies between the Discrete Gravitational Ontology and Artificial Neural Systems

Description

Abstract

The Discrete Gravitational Ontology (DGO) proposes that spacetime, matter, and quantum
behavior emerge from an evolving network of discrete gravitational relations. Each relation
represents a minimal act of causal connectivity, and the universe advances through successive
relational updates constrained by an invariant rate c = ℓ0/τ0. This paper presents a conceptual
and methodological exploration rather than an empirical study. It examines a formal analogy
between the dynamics of DGO and the architecture of artificial neural networks (ANNs). Both
systems consist of nodes connected by weighted relations that evolve through local update rules
minimizing a global inconsistency measure—the DGO action S[G] and the ANN loss function
L(W), respectively. In this analogy, spacetime geometry corresponds to a learned representation
of relational coherence, while the irreversible arrow of time parallels the unidirectional optimization
flow of learning. The paper argues that interpreting the universe as a self-learning relational
network offers a coherent framework linking physical law and informational dynamics. The goal
is not to propose new empirical predictions but to outline a conceptual foundation that could
inspire future theoretical and computational research on the emergence of geometry, information,
and intelligence from discrete relational principles. The paper also serves as a structured
record of the author’s ongoing research and learning process.