Skip to content

Xitral/beacon-space-ai

Repository files navigation

BEACON

CI LaTeX Paper Version DOI Concept DOI Python License Status Release Candidate

BEACON (Bayesian Event Assessment for Conjunction Observation and Notification) is a reproducible research project for calibrated, uncertainty-aware satellite conjunction triage using public CDM data.

BEACON is a research prototype only, not an operational system.

Current development version: v0.3.0-rc1. The latest archived reproducible artifact is v0.2.2; v0.3.0-rc1 is a release candidate and does not yet have a final Zenodo version DOI.

Citation and Archived Release

Version DOI: 10.5281/zenodo.21209794
Concept DOI: 10.5281/zenodo.21209119
Archived version: v0.2.2
Current release candidate: v0.3.0-rc1
Repository: https://github.com/Xitral/beacon-space-ai

Use the version DOI to cite the exact v0.2.2 artifact used for reproducibility. Use the concept DOI to cite the overall BEACON archive across versions. If you use BEACON, also cite the original public dataset provider.

License

BEACON source code on the current main branch is released under the Apache License 2.0. See LICENSE for details.

Licensing note: BEACON versions prior to v0.3.0-rc1, including the archived v0.2.2 Zenodo version DOI listed above, were released under the MIT License. Beginning with v0.3.0-rc1 and future releases, BEACON source code is released under the Apache License 2.0.

Documentation, figures, research materials, and demo scenarios are licensed under CC BY 4.0 unless otherwise noted. BEACON names, logos, and brand assets are not open-licensed unless explicitly marked.

Research Focus

BEACON evaluates rare-event ranking, probability calibration, uncertainty-aware human review, repeated split robustness, current-risk feature ablation, leakage-safe evaluation, and interactive visual analytics for model-grounded triage inspection.

Reproducibility

python -m pip install -r requirements.txt
python src/run_all.py
python -m pytest -q

Raw challenge data is not committed to the repository. See data/README.md for expected filenames and required columns.

Documentation

  • REPRODUCIBILITY.md defines the environment, commands, expected outputs, expected metric ranges, viewer checks, paper build steps, and caveats.
  • docs/release_validation_v0.3.0.md records the current release-candidate validation status.
  • docs/release_notes_v0.3.0.md contains draft release notes.
  • docs/reviewer_summary_v0.3.0.md provides a short reviewer-facing summary.
  • docs/viewer_demo_checklist.md provides manual viewer QA and export validation steps.

Paper

The manuscript source is available in paper/main.md and paper/main.tex.

About

Trustworthy AI for satellite conjunction triage using calibrated ML, Bayesian logistic regression, and uncertainty-aware escalation on public CDM data.

Topics

Resources

License

Stars

1 star

Watchers

1 watching

Forks

Packages

 
 
 

Contributors