Skip to content

README.md

Latest commit

 

History

History
28 lines (21 loc) · 2.01 KB

README.md

File metadata and controls

28 lines (21 loc) · 2.01 KB

GNN Implementations Documentation

Version: v1.6.0 Engine (Bundle v2.0.0)
Last Updated: 2026-04-14
Status: ✅ Production Ready
Modules: 38+ · Pipeline steps: 25 · Renderers: 9 backend guides in this directory · Tests: see ../../../README.md

This directory contains documentation and references for the Implementations domain of Generalized Notation Notation (GNN).

Available Documents

  • PyMDP: The canonical reference implementation for discrete True POMDP simulation (pymdp). Reached 1.0 Correlation Baseline.
  • NumPyro: Probabilistic programming for continuous distributions, uncertainty mechanics, and MCMC/SVI (numpyro >= 0.14). Verified as Fully Operational in March 2026.
  • PyTorch: Neural Active Inference with learnable parameters, differentiable gradients, and GPU acceleration (torch >= 2.0). Verified as Fully Operational in March 2026.
  • JAX: High-performance numerical computing and XLA vector-space compilation (jax). Reached 1.0 Correlation Baseline.
  • RxInfer.jl: Reactive message passing and declarative probabilistic programming in Julia (RxInfer.jl). Reached 1.0 Correlation Baseline.
  • ActiveInference.jl: Dedicated discrete-state Active Inference simulation in Julia (ActiveInference.jl).
  • DisCoPy: Categorical string diagrams enabling advanced symmetry representations and compositional verification semantics for Multi-Agent Topologies (discopy).
  • CatColab: Topos Institute framework mapping structural logic into Schema and Ologs, acting as a categorical twin.
  • Stan: Stan model code generation from GNN variables and connections (structural heuristics; no full Active Inference semantics).

Navigation

GNN: A text-based language for Active Inference generative models.