Can a neural network trained on large volumes of simulated bearings-only data produce usable target range, course and speed estimates? This repository is a TypeScript proof of concept that answers that question through a series of go/no-go gates. It is not a product.
- Site: https://debrief.github.io/Gitma/ (gate status, blog, glossary, data dictionary, algorithms)
- Requirements:
docs/SRD.md - Principles:
.specify/memory/constitution.md - Current work:
specs/(Spec Kit). Walkthroughs: iteration 0quickstart.md, iteration 1quickstart.md(training,tma-g1) - Contributing / agents:
AGENTS.md
| Gate | Result | Evidence |
|---|---|---|
| G0 — pipeline verified | pass (agent review panel, round 3) | experiments/20261011-g0-r3 |
| G1 — signal exists | training (iteration 1, GitHub Actions); panel review to follow | specs/002-iteration-1-signal |
Requires Node 22 (.nvmrc), pnpm and uv.
pnpm install && uv sync && pnpm build
pnpm test
pnpm tma-generate --seed 42 --count 100000 --noise-free --out data/g0
pnpm tma-g0 --data data/g0 --experiment experiments/my-g0
All data is simulated; nothing above OFFICIAL belongs in this repository.