Digital Transformer for Fracture Classification: Part I. SYM-Fractron, A Physics-Driven Hybrid Binary–Symbolic Computing Ecosystem

Fracture networks strongly control fluid flow, reservoir connectivity, and production performance in carbonate systems, yet their multiscale architecture of complexity remains difficult to characterize from heterogeneous geological and geophysical datasets. Here, we introduce the Digital Transformer (DiT), a physics-informed computational framework that automatically analyzes and classifies fracture systems using spatially encoded visuonumerical primitives derived directly from physical measurements. Instead of relying on textual tokenization, the approach performs attention primitives tokenization of multiscale geophysical data. Clusters of absolute integer values act as computational tokens while preserving spatial topology and scale-invariant structure of the original system. The framework integrates two complementary environments: Muuk'il Kaab (MIK) for multidimensional metadata fusion and visualization, and SYM-Fractron, a hybrid binary-symbolic transformer for two-dimensional image analysis. Within this architecture, Digital Twins provide coupled visual and statistical representations of geological systems and their computational counterparts, enabling an interpretable taxonomy of natural fracture patterns while supporting well-trajectory optimization in the exploration of dolomitized carbonate reservoirs. In this view, fracture architectures become visionumerical primitives whose physics-informed tokenization opens a pathway from the architecture of natural complexity to its computational realization through Digital Twins.