[ OK ] /methods :: how Artifact-SETI searches [ OK ] approach ........................................ conventional explanations first [ OK ] data sources .................................... 7 catalogs / archives [ OK ] inquiry types ................................... 4 modes [ ON ] apparatus visible.
Methods
─────────────────────────────────────────────────────────────────────────
[ RESEARCH BASIS ]
Bracewell's autonomous messenger probe, Papagiannis's proposal to examine the asteroid belt, and Benford's observational program for nearby co-orbital “lurkers” define the working hypothesis: durable extraterrestrial technology could be a local astronomical observable. They motivate where to look; they do not license an artificial interpretation of any anomaly. Full citations are listed in research notes.
────────────────────────────────────────────────────────────────────────────
[ SEARCH APPROACH ]
Artifact-SETI uses the possibility of extraterrestrial artifacts to motivate testable searches. Natural, instrumental, statistical, and human explanations remain the starting points for interpreting an outlier. A residual is a discrepancy relative to a particular model, not evidence that conventional explanations are exhausted.
These are checks to document, including their limits and any unfinished work:
candidate ─▸ catalog ─▸ dynamics ─▸ sensor ─▸ selection ─▸ pipeline ─▸ residual A failed check may downgrade a candidate; a passed check addresses only the explanation and assumptions tested. Inconclusive and missing checks remain visible.
─────────────────────────────────────────────────────────────────────────
[ DATA SOURCES ]
The pilots use public archives and catalogues. Public data access alone does not ensure reproducibility: exact queries, versions, selections, code, and intermediate results are also needed. The sources below describe available inputs; coverage differs by inquiry.
| Source | Family | What it gives us | Program use |
|---|---|---|---|
| JPL SBDB | small bodies | Orbital elements and physical parameters for asteroids, comets, NEOs. | candidate selection · orbit lookup |
| JPL Horizons | ephemerides | High-precision state vectors and ephemerides for any catalog body. | observation planning · model context |
| IRSA / NEOWISE | thermal IR | WISE / NEOWISE imaging and reactivation photometry. | albedo + size · thermal proxies |
| VizieR | aggregator | Cross-matched access to thousands of astronomy catalogs. | cross-survey residuals · prior literature |
| Gaia | astrometry | Sub-mas astrometry, parallax, photometry for ~1.8B sources. | astrometric backbone · stellar context |
| TESS / ATLAS | time domain | Optical light curves and time-domain observations. | occultation candidates · time series |
| MAST / HST | archive | Hubble + Kepler + JWST archive imaging and spectra. | archival imaging · counterpart checks |
A catalogue non-match is informative only within the checked coverage and detection limits. Several catalogues may share observations or processing assumptions.
─────────────────────────────────────────────────────────────────────────
[ INQUIRY TYPES ]
The four pilot inquiries use different screening procedures and internal labels. Scores and categories are not directly comparable between inquiries.
Compare several physical and observational dimensions.
Scores across orbital, thermal, reflectance, size, and variability measurements prioritize review. Shared inputs and correlated errors must be assessed before treating their agreement as corroboration.
EXAMPLESJupiter Trojan multi-axis survey
Test unusual orbital populations against selection and dynamics.
Catalogue counts describe the selected sample. A physical population claim additionally requires uncertainty analysis, natural dynamical models, and survey selection and completeness estimates.
EXAMPLESUltra-circular near-1 au population
Follow a moving target through independent public archives.
Ephemerides are joined to optical, ultraviolet, infrared, and radio coverage so candidate behavior can be checked across epochs, instruments, and wavelengths.
EXAMPLESKuiper Belt / TNO optical anomaly hunt
Seek corroboration in additional measurements.
Additional surveys can help test a residual when coverage and sensitivity are adequate. A cross-survey filter label alone is not independent confirmation.
EXAMPLESMain-belt cross-survey anomaly search
─────────────────────────────────────────────────────────────────────────
[ EVIDENCE AND REPRODUCIBILITY ]
The website summarizes exploratory pilots. Internal labels such as P, “strict residual,” and the legacy “artifact score” describe screening decisions, not validated probabilities of artificiality. These summaries should not be read as preregistered confirmation or independent replication.
A null result constrains only signatures the search could detect. Known controls, injected signals, false-positive estimates, and search-wide error control are needed to interpret both positive and negative outcomes. Agreement among AI agents or reruns on the same data does not supply independent empirical evidence.
For a confirmatory follow-up, freeze the hypothesis, data selection, thresholds, stopping rules, and adjudication procedure before testing on new or genuinely held-out data. A reproducible package should preserve code, prompts, model versions, data provenance, and audit records, with expert review of the design and escalations.
The broader question is whether agentic AI makes these searches more affordable. That requires comparing workflows at similar scientific quality and counting human supervision and review alongside computation. The pilots do not yet establish those cost savings.
Interactive views are summaries of stored results, not live archive feeds. Schematic orbital views illustrate geometry, not measured positions or trajectories. Downloadable PDFs preserve earlier reports and may use stronger language than the current web summaries; the limitations on each inquiry page apply.