Void Dynamics Model (VDM)
A two part framework built to be attacked:
A background‑independent, metriplectic field theory unifying matter and forces from the Quantum Geometric Tensor, with an emergent causal cone and an epistemological J→M projection.
A real time, zero training, emergent run-time with scale‑free/heavy‑tail sparse neural graphs, locally interacting neuron‑particles (void walkers), Hebbian working memory/online plasticity, sparse activations, and attention‑like state routing driven by physical dynamics.
Lineage: AMN-->FUM-->VDM
🔭 What I’m building
- VDM — Klein-Gordon / reaction–diffusion field intelligence + walker ecology + scoreboard/GDSP gating
- Brain-Like Cognition — emergent cognition self-organizes from strict fundamental physics constraints and avoidance of traditional ML tactics
- Memory Steering — dynamic knowledge graph with event-driven updates
- Agency — inherent, minimal energy optimization rules
- Real-time control — swap massive pretraining for fast constraint satisfaction
Reproducibility: baselines + QA artifacts are archived on Zenodo. Code lives in public GitHub repos with tests and docs.
🧪 Reproducible records (Zenodo)
- Latest Zenodo Uploads: Void Dynamics Model: Zenodo Community
- (with code + datasets + figures)