Scientific assurance
Methods for VVUQ, assurance on the Julia and Python and r and rust compute planes.
45 capabilities
16 of 45 have a bound kernel or engine. States are shown as the registry records them; a request fails closed wherever an adapter, specialist model or governed data is still missing.
| Method | Capability ID | State | Runs on |
|---|---|---|---|
| Unit Verification | risk.capability.scientific-assurance.unit.verification.v1 |
Reference implementation | – |
| Analytical Benchmark | risk.capability.scientific-assurance.analytical.benchmark.v1 |
Adapter required | Python · validation.error_metrics |
| Convergence Testing | risk.capability.scientific-assurance.convergence.testing.v1 |
Adapter required | Python · validation.convergence_order |
| Numerical Stability | risk.capability.scientific-assurance.numerical.stability.v1 |
Adapter required | Python · validation.invariants |
| Solver Tolerance | risk.capability.scientific-assurance.solver.tolerance.v1 |
Adapter required | – |
| Mesh Grid Convergence | risk.capability.scientific-assurance.mesh.grid.convergence.v1 |
Adapter required | – |
| Reproducibility | risk.capability.scientific-assurance.reproducibility.v1 |
Reference implementation | Python · validation.replay_compare |
| Cross Language Parity | risk.capability.scientific-assurance.cross.language.parity.v1 |
Adapter required | Python · validation.parity_compare |
| Deterministic Replay | risk.capability.scientific-assurance.deterministic.replay.v1 |
Reference implementation | – |
| Calibration | risk.capability.scientific-assurance.calibration.v1 |
Adapter required | Python · validation.reliability_diagram |
| Holdout Testing | risk.capability.scientific-assurance.holdout.testing.v1 |
Adapter required | Python · validation.error_metrics |
| Cross Validation | risk.capability.scientific-assurance.cross.validation.v1 |
Adapter required | Python · spatial_validation.blocked_metrics |
| Hindcasting | risk.capability.scientific-assurance.hindcasting.v1 |
Adapter required | – |
| Backtesting | risk.capability.scientific-assurance.backtesting.v1 |
Adapter required | Python · validation.error_metrics |
| Event Reconstruction | risk.capability.scientific-assurance.event.reconstruction.v1 |
Adapter required | – |
| Reliability Diagram | risk.capability.scientific-assurance.reliability.diagram.v1 |
Adapter required | Python · validation.reliability_diagram |
| Residual Diagnostics | risk.capability.scientific-assurance.residual.diagnostics.v1 |
Adapter required | – |
| Goodness of Fit | risk.capability.scientific-assurance.goodness.of.fit.v1 |
Adapter required | – |
| Predictive Scoring | risk.capability.scientific-assurance.predictive.scoring.v1 |
Reference implementation | Julia · frontier_risk_math.brier_score |
| External Benchmark | risk.capability.scientific-assurance.external.benchmark.v1 |
Adapter required | – |
| Independent Replication | risk.capability.scientific-assurance.independent.replication.v1 |
Adapter required | – |
| Aleatory Uncertainty | risk.capability.scientific-assurance.aleatory.uncertainty.v1 |
Adapter required | – |
| Epistemic Uncertainty | risk.capability.scientific-assurance.epistemic.uncertainty.v1 |
Adapter required | – |
| Parameter Uncertainty | risk.capability.scientific-assurance.parameter.uncertainty.v1 |
Adapter required | – |
| Model Form Uncertainty | risk.capability.scientific-assurance.model.form.uncertainty.v1 |
Adapter required | – |
| Data Uncertainty | risk.capability.scientific-assurance.data.uncertainty.v1 |
Adapter required | – |
| Scenario Uncertainty | risk.capability.scientific-assurance.scenario.uncertainty.v1 |
Adapter required | – |
| Structural Uncertainty | risk.capability.scientific-assurance.structural.uncertainty.v1 |
Adapter required | – |
| Measurement Uncertainty | risk.capability.scientific-assurance.measurement.uncertainty.v1 |
Adapter required | – |
| Spatial Uncertainty | risk.capability.scientific-assurance.spatial.uncertainty.v1 |
Adapter required | – |
| Temporal Uncertainty | risk.capability.scientific-assurance.temporal.uncertainty.v1 |
Adapter required | – |
| Applicability Domain | risk.capability.scientific-assurance.applicability.domain.v1 |
Connected | engine.grix.risk.validation.vvuq.applicability.v1 |
| Geographic Applicability | risk.capability.scientific-assurance.geographic.applicability.v1 |
Adapter required | – |
| Population Applicability | risk.capability.scientific-assurance.population.applicability.v1 |
Adapter required | – |
| Hazard Regime Applicability | risk.capability.scientific-assurance.hazard.regime.applicability.v1 |
Adapter required | – |
| Asset Class Applicability | risk.capability.scientific-assurance.asset.class.applicability.v1 |
Adapter required | – |
| Temporal Horizon Applicability | risk.capability.scientific-assurance.temporal.horizon.applicability.v1 |
Adapter required | – |
| Scenario Applicability | risk.capability.scientific-assurance.scenario.applicability.v1 |
Adapter required | – |
| Resolution Applicability | risk.capability.scientific-assurance.resolution.applicability.v1 |
Adapter required | – |
| Data Quality Applicability | risk.capability.scientific-assurance.data.quality.applicability.v1 |
Adapter required | – |
| Intended Use Applicability | risk.capability.scientific-assurance.intended.use.applicability.v1 |
Adapter required | – |
| Categorical Forecast Verification | risk.capability.scientific-assurance.categorical.forecast.verification.v1 |
Adapter required | Python · verification.contingency |
| Probabilistic Forecast Skill | risk.capability.scientific-assurance.probabilistic.forecast.skill.v1 |
Adapter required | Python · verification.brier_skill |
| ROC Analysis | risk.capability.scientific-assurance.roc.analysis.v1 |
Adapter required | Python · verification.roc |
| CRPS Verification | risk.capability.scientific-assurance.crps.verification.v1 |
Adapter required | Python · verification.crps |
Methods, data and supports
- Method families
- VVUQ, robust statistics, Bayesian inference, global sensitivity
- Domains
- VVUQ, assurance
- Compute planes
- Julia, Python, r, rust
- Source classes
- benchmark, calibration, holdout, run-receipt, external-review
- Spatial supports
- model-domain
- Temporal supports
- model-version, evaluation-window
Propose a method
New methods are registered with a stable ID and a hand-derived golden test.
অবস্থা
The methodology, structure and definitions of The Global Risks Index are a preparatory working draft (2026), circulated for review and comment and not for implementation. No value appears until an edition is released from a governed evidence cut; values then carry intervals, classes, sources and their correction pathway. Simulators run the published methods on numbers you choose and never show a NXS-GRX value.
NXS-GRX যা নয়
NXS-GRX গবেষণা ও ঝুঁকি-বিশ্লেষণ তথ্য। এটি ঋণ বা ESG রেটিং নয় এবং কোনো রেটিং স্কেল ব্যবহার করে না; এটি আর্থিক উপকরণের বেঞ্চমার্ক নয়, সরকারি পরিসংখ্যান নয়, জরুরি সতর্কবার্তা বা সরকারি শ্রেণিবিভাগ নয় এবং পরামর্শ নয়। প্রাকৃতিক ব্যক্তিদের মূল্যায়নে এটি ব্যবহার করা যাবে না। ঝুঁকির শ্রেণি হলো একটি প্রকাশিত স্কেলের ব্যান্ড, কখনো দেশ বা মানুষের র্যাংকিং নয়।
প্রতিষ্ঠান ও উৎস
The Nexus Consortium is stewarded by three independent not-for-profit institutions: The Global Centre for Risk and Innovation (GCRI), through GCRI US in Washington, DC and GCRI Canada in Ontario, with GCRI Canada as Lead Institution, Consortium Secretariat and lead contracting organisation; The Global Risks Forum (GRF); and The Global Risks Alliance (GRA). Technology, infrastructure and service providers work with the Consortium at arm’s length under written agreements, like every other external party in the public-private-planet model, and hold no role in its stewardship.
Every value carries the attribution of its sources, which keep their own licences. External names are cited for interoperability only. Nexus-authored text is intended for CC BY 4.0 on release.
Avis en français
La méthodologie et les définitions du NXS-GRX sont un projet de travail préparatoire (2026). Le NXS-GRX est une information de recherche et d’analyse des risques : ni notation, ni indice de référence financier, ni statistique officielle, ni alerte, ni conseil. Les classes de risque sont des bandes sur une échelle publiée, non des classements.