Boundaries
Writing non-goals is part of credibility.
Use boundaries, murine-first rationale, Verification-lite positioning, and the “pending public-data runs” results policy. The medical disclaimer in the footer always remains.
§ 1
Non-goals (v1)
- ×
Individual-patient digital twins
No personal prediction or scheduling.
- ×
Medical advice / clinical decisions
Outputs for research demos, teaching and early mechanism studies only.
- ×
Molecule generation
v1 AI searches intervention strategies only.
- ×
Subjective competitor scores / named bake-offs
Not published on this site.
- ×
Fabricated run results
No fit errors, measured coverage or pass/fail before public-data runs.
§ 2
Why murine-first
Richer controllable time series; lower heterogeneity; easier L2; more auditable observables. v1 builds a murine credibility baseline; human-data extrapolation is a v2+ discussion.
§ 3
Verification-lite positioning
An engineering floor for dt/grid sensitivity; default key-endpoint change ≤ 20% (COU-adjustable). Explicitly not full numerical verification.
§ 4
Results publication policy
All result-class numbers: pending public-data runs. DOI TBD. Paper status: in preparation.
§ 5
Label legend
| Label | Meaning |
|---|---|
| Planned | Not delivered yet |
| Default acceptance threshold | Design target; COU-adjustable; not measured |
| Schematic | Not model-run output |
| Structure example | Field sketch; explanatory values |
| Replaceable | Default spine can change |
| Pending public-data runs | Results / errors / pass-fail |