Checklist
Interactive Tool

Model Validation & Reproducibility Checklist

Use this interactive checklist to capture evidence that your AI or ML model meets your research's reproducibility and validation needs. Save answers so you can attach them to manuscripts, audits, or deployment records.

Who collected the data, when, and what preprocessing (filters, exclusions) was applied?
Include splitting method (random, stratified, time-based) and random seeds to allow exact replication.
Baselines may include mean predictor, logistic regression, heuristic rules, or domain-specific heuristics.
List metrics, bootstrap/CIs if used, and code or notebooks used to compute them.
E.g., prediction intervals, calibration plots, or Bayesian posteriors.
External validation reduces overfitting risk. Note dataset name and differences from training data.
Methods include SHAP, LIME, attention maps, or domain-specific sensitivity analyses.
Pre-registration reduces p-hacking and post-hoc tuning. Provide link or file if available.
Model Cards concisely communicate who should and should not use a model and known failure modes.
Record repository links, commit hashes, and exact model-artifact locations (e.g., artifact store path).
List smoke tests and acceptable thresholds before the model is used in practice.
Include metrics to monitor, alert thresholds, and a rollback plan.
Attach approvals or note pending reviews.
You can explore this tool now. Sign in or create an account to save your responses and return to them later.
Sign In to Save Create an Account
Make this tool part of your work

Save a personal copy, bring it to your team, or tailor the questions and workflow to fit what you are hungry to improve.

Member customization and team collaboration are coming soon.
Discuss Ask the Librarian

Discussion

Comments and conversation will live here.