WorldEnv
WorldEnv is an OpenEnv environment factory powered by canvas-engineering and a General Unified World Model (GUWM).
Core trick
Heterogeneous datasets train one backbone via missing-field masking, so loss is computed only on fields with ground truth.
Description
WorldEnv is an OpenEnv environment factory powered by canvas-engineering and a General Unified World Model (GUWM). Canvas-engineering provides a type system for multimodal latent computation: typed regions on a 3D spatiotemporal grid define fields with their own frequency, loss weight, role, and semantic embedding, while a CanvasTopology defines the attention compute graph. This allows heterogeneous datasets to train a shared backbone using missing-field masking—loss is computed only where ground truth exists, enabling transfer across domains without imputation.
The GUWM defines an 857-field causal ontology spanning 19 layers (physics, markets, macroeconomy, firms, individuals, events, etc.). Any subset of fields can be projected onto a canvas to instantiate an RL environment. Corporate strategy, robotics, or macroeconomic environments are simply different projections of the same world model, enabling shared representations and cross-domain transfer.
JacobFV/general-unified-world-modeling★ 10 forksPythonNOASSERTIONCreated Mar 8, 2026 · last push Mar 8, 2026
README excerpt
# general-unified-world-model ### A typed causal ontology of civilization, built on [canvas-engineering](https://github.com/JacobFV/canvas-engineering) structured latent spaces. [](https://pypi.org/project/general-unified-world-model/) [](https://github.com/JacobFV/general-unified-world-modeling/actions/workflows/ci.yml) []() [](https://www.python.org/downloads/) [](LICENSE) > Canvas engineering structures what a diffusion model *thinks in*. This repo declares a **857-field typed schema** spanning planetary physics through individual psychology, compiles it onto a structured latent canvas, and trains it on heterogeneous real-world data — without throwing out samples that are missing fields. --- <!-- Figure: Heatmap of the full 857-field world model packed onto a canvas grid. Each rectangular region is color-coded by semantic domain (financial=teal, macro=blue, regime=red, narratives=purple, etc.). The spatial layout shows how fields are strip-packed left-to-right, top-to-bottom, with larger domains (financial, country blocks) occupying more area. --> <p align="center"> <img src="https://raw.githubuser