MJos Predictive Model | First manifestation of the Bangel Language
New project sequence ยท proposed bindings
Associative recall, relational discrimination and bounded source-grounded diagram overlays.
MJ Neural Net / Indigo MJos Predictive Model | First manifestation of the Bangel Language MJ Physics Engineering | Owned and operated by Michael Bangel | Clearwater, Florida VERIFICATION AND DISCOVERY RECORD Project: MJ Neural Net OBSERVED IN THIS PASS - Source defines nine-node activations, Hebbian recurrent weights, retrieval, relational filter and bounded command placement. - Three nine-node associative layers; five recovered diagram-overlay templates retained under a separate envelope. - Source command-to-node model maps nine Bangel commands and predicts first-order transitions. EVIDENCE STATUS Source/document inspection only for this project sequence. No historical test count is restated as a fresh pass. No fresh project runtime suite was executed by authoring this paper. Studio starter checks and library integrity checks, where completed, have their own separately scoped records. NEXT EXECUTABLE ASSERTIONS Check that admitted inputs produce the stated bounded result; missing inputs HOLD; failed permission DENY; stale/duplicate identities do not create a second effect; successful calculation creates no unauthorized action. Expected outcomes must be fixed before running. RESULT PROVENANCE Record test identity, exact source revision, command, environment, expected/actual outcome and transcript hash. Native test code and result evidence remain authoritative. This is a new reconstruction based on the recovered Indigo rules. EVIDENCE LIMITS - Engineered computational model, not hippocampal biology or general natural-language prediction. - Lowercase r0 remains provisional and distinct from uppercase R0. - Synthetic comparison shows a recall/coverage tradeoff; not field accuracy or equal-coverage benefit. - Diagram geometry and command mapping must retain separate namespaces.