d0m8n.io
MnΛ Corp.US · Two Door Solutions · Dayton, Ohio
What this is
Independent research in coherence architectures, polymodal compression notation, and structural identity theory. Solo work. Not venture-backed. Not a startup pitch. Just ideas that I think matter, built carefully over time.
Some of this is elegant. None of it has empirical benchmarks yet. I'm being honest about that because the ideas deserve honest framing.
Active work
GhostNode v1.5
Directed-graph coherence engine. Tracks structural identity across state transitions using a triple-observer system. NumPy/NetworkX substrate. The core question: can you measure whether a system is still itself after transformation?
Working prototype
Compr1sion
Identity-preserving recursive inference. A compression principle where the compressed form retains structural identity with the original — not just information, but coherence. Formal spec complete. Semantic input adapter and decode module still needed for MVP.
Research prototype
Polymodal compression notation
A notation system where tokens fire simultaneously across orthographic, phonetic, semantic, and structural registers. Started as a personal expressive mode. Now being formalized into a machine-readable compression language. Corpus captured, taxonomy in progress.
Active development
Structural identity argument
The claim that structural identity invariants differ fundamentally from rule-based guardrails — because violating them causes measurable self-collapse, not just policy violation. Developed as a framework for AI alignment. Needs empirical validation.
Theoretical
AiQ² / QueRe
Query Recursive Engine with quaternion state control. Runtime-level contextual inference governance across a 15-layer architecture. Core engine and API surface deployed locally. Not yet production-hardened.
Local deployment
This work is at various stages of completion. The ideas are original and, I believe, important. What's missing is empirical benchmarking — the data that turns a coherent theory into a proven system. That's the next phase. If you're interested in the research, in collaboration, or in the problems these tools are trying to solve, reach out.