Pattana-Relational Dynamics (PRD) Causal AI Engine v1.8 presents a multi-phase framework integrating relational algebra, causal reasoning, neural architectures, and reinforcement-based decision systems. The project develops an AI system grounded in a 24-dimensional relational operator structure derived from SU(5) algebra, where causal relations are represented as relational states evolving through neural and graph-based mechanisms.The framework progresses through six phases, beginning with the algebraic definition of relational operators and extending to autonomous causal reasoning agents capable of counterfactual simulation and reinforcement-driven policy optimization.Phase 1 establishes the algebraic foundation by defining the 24 relational generators and their commutation relations, forming the core relational state space. Phase 2 introduces relational neural network layers that map relational state vectors to causal influence outputs. Phase 3 extends this representation through multi-step causal graph propagation, allowing inference across relational dependencies.Phase 4 incorporates temporal recurrent dynamics, enabling the system to model evolving causal structures over time. Phase 5 introduces counterfactual reasoning and intervention modeling based on causal graph modifications inspired by intervention calculus. Phase 6 completes the architecture with reinforcement-driven autonomous decision loops, enabling the system to select optimal interventions through reward-guided policy optimization.The resulting architecture integrates relational algebra, neural networks, causal graphs, and reinforcement learning into a unified causal AI framework. This repository contains the complete LaTeX documentation for Phases 1–6 along with prototype implementation structures intended for further development of causal reasoning AI systems.The release is published as version 1.8 under a single DOI to maintain continuity across all phases of the PRD Causal AI Engine framework.
Myomin Aung (Tue,) studied this question.