{ "id": "AIBD-0283", "title": "MADGRAV anomaly-detection pipeline recovers 47 gravitational-wave detections in LIGO O3 and O4 data, 44 shared with coherent WaveBurst and none new", "summary": "Researchers used MADGRAV-2D to recover 47 cataloged gravitational-wave signals from LIGO data.", "date_observed": "2026-09-30", "date_basis": "source publication", "date_precision": "day", "benefit_status": null, "source_collection": null, "source_entry": null, "account": "A paper was posted on arXiv on 30 September 2026 by Gianluca Inguglia, Huw Haigh and Ulyana Dupletsa. They are at the Marietta Blau Institute for Particle Physics (Austrian Academy of Sciences, Vienna). Alessandro Longo (University of Urbino and INFN Florence) is a co-author. It is titled “MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data” (version 2, 19 October 2026; submitted to the European Physical Journal C). The pipeline is “a series of sequential convolutional neural networks that perform anomaly detection, glitch classification, coherence testing, and signal ranking”. It is applied to one-second Q-transform windows of Hanford and Livingston data from O3a, O3b, O4a and O4b. That is 442.9 days of coincident livetime in total. The abstract reports 47 detections with a false alarm rate below one per year and p_astro above 0.9. Of those, “44 are shared with the minimally modeled coherent WaveBurst search”. The three others are cataloged events, and the paper states that “no candidate without a catalog counterpart survives the analysis”. Recovery rises with mass. It reaches 8.1% of confident catalog events below 30 solar masses and 39.8% between 30 and 100. It reaches 53.3% above 100 for network SNR above 10. The authors conclude that anomaly-detection pipelines “can serve as an independent detection channel complementary to matched filtering in the high-mass high-SNR regime”. The paper confirms no new event independently. The authors identify the two-detector configuration as MADGRAV-2D in their code repository. Its changelog dates the final pipeline configuration to 14 September 2026, and its README states that this configuration produced the reported 47 detections. The repository provides the trained weights and frozen calibration.", "category": "general benefit", "unexpected": false, "purpose": "MADGRAV was built to search LIGO data for high-mass compact binary coalescences by anomaly detection. Recovering cataloged events was that purpose.", "system": "MADGRAV-2D (two-detector cascade of convolutional neural networks)", "model_version": null, "developer": "Marietta Blau Institute for Particle Physics, Austrian Academy of Sciences, with the University of Urbino and INFN Florence", "deployer": "The same authors, on public LIGO O3 and O4 data", "location": "Vienna, Austria", "domain": "Astrophysics / gravitational-wave astronomy", "segment": "Academic research on public detector data", "benefit_type": "Improved Task Accuracy", "taxonomy_note": "Re-detection of cataloged events is read as task performance; no discovery is claimed.", "registry_ids": {}, "ratings": [ { "group": "organizational", "depth": null, "reason": "Recovery of 47 known events, 44 of them shared with coherent WaveBurst, demonstrates the method. No new event, catalog adoption or operational use is documented." }, { "group": "societal", "depth": null, "reason": "A complementary search channel is proposed; no scientific result beyond re-detection is claimed." } ], "ai_role": { "status": "assessed", "statement": "The convolutional networks performed the whole detection chain, from anomaly detection to ranking, and the authors are the developers. Every detection matches a published catalog event. The authors state that no candidate without a catalog counterpart survives, so the search added no new discovery in this analysis." }, "evidence": [ { "kind": "preprint", "citation": "Inguglia G, Haigh H, Dupletsa U, Longo A (2026). MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data. arXiv:2609.39583v1, 30 September 2026, abstract and sections 3–5. Revised as v2 on 19 October 2026. Authored by the pipeline's developers.", "title": "MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data", "url": "https://arxiv.org/abs/2609.39583v1", "date": "2026-09-30", "written_by": "makers", "quote": "Of the 47 detections, 44 are shared with the minimally modelled coherent WaveBurst search." }, { "kind": "other", "citation": "Inguglia, Gianluca, and collaborators. MADGRAV code repository, snapshot acb66841f9408939c1ca17bff90360214c4ea2c9, 19 September 2026. README opening and sections 3–6; CHANGELOG entry for 14 September 2026; spectrogram_cascade/massive_calibration_BA.json; environment.yml.", "title": "MADGRAV: Multilevel Anomaly Detection for GRAVitational-wave science", "url": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "date": "2026-09-19", "written_by": "makers", "quote": "This is the final version of the pipeline" } ], "replication": { "produced_by": [ "Sequential CNNs for anomaly detection, glitch classification, coherence testing and signal ranking on one-second Q-transform windows of Hanford and Livingston data", "442.9 days of coincident livetime across O3a, O3b, O4a and O4b" ], "must_hold": [ "High-mass, high-SNR sources: the recovered fraction rises from 8.1% below 30 solar masses to 53.3% above 100", "Two-detector coherence between Hanford and Livingston" ], "echoes": [] }, "paired_incident": null, "headline": "Researchers found 47 known gravitational-wave signals", "metric": "47 detections with a false alarm rate below one per year and p_astro above 0.9", "comparator": { "against": "the coherent WaveBurst (cWB) search and the published catalogs on the same LIGO data", "before": null, "after": "47 detections, 17% of the catalog population and 39% of the cWB count", "basis": "measured", "quote": "MADGRAV 47 detections correspond to 17% of the catalog population and 39% of the cWB count. 44 of the 47 are shared with cWB: the MADGRAV detection set is a subset of cWB, with the exclusion of 3 events.", "url": "https://arxiv.org/html/2609.39583v2" }, "later": { "status": "unknown", "date": null, "note": "The authors revised the preprint on 19 October 2026. No independent follow-up result is cited.", "quote": null, "url": null }, "value": null, "developer_country": "Austria", "deployer_country": "Austria", "unrated_because": null, "brief_notes": null, "brief": null, "revisions": [ { "date": "2026-10-20", "change": "Record drafted from the cited sources.", "source": null, "editor": "AIBD" }, { "date": "2026-10-21", "change": "Title reworded to state the event without naming who claims it.", "source": null, "editor": "AIBD" }, { "date": "2026-10-21", "change": "Added what the benefit is measured against from the cited sources.", "source": "https://arxiv.org/html/2609.39583v2", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Benefit type set to Improved Task Accuracy under Trustible's full taxonomy of 23 types; it was Improved Knowledge Discovery among the six first used.", "source": "https://trustible.ai/post/a-framework-for-measuring-the-benefits-of-ai/", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the developer and the deployer added from the organization's own site or a registry.", "source": "https://inspirehep.net/api/institutions?q=Marietta%20Blau&size=3; https://inspirehep.net/api/institutions?q=Marietta%20Blau&size=3", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Prose pass: a rating reason, account, purpose, the system's part rewritten in shorter sentences with abbreviations expanded; facts, figures and quotations unchanged.", "source": null, "editor": "AIBD" }, { "date": "2026-10-21", "change": "Notes about which sources could be read were reworded for readers; facts, figures and quotations unchanged.", "source": null, "editor": "AIBD" }, { "date": "2026-10-24", "change": "Consolidated the abstract and HTML citations into one primary preprint. Read version 1 and verified that the 47 detections, 44 shared detections, 442.9 days of data, recovery rates, comparison percentages, and absence of a new candidate were already reported on 30 September 2026. Preserved the 19 October 2026 version 2 link and the submission history in source provenance. Versions of the same work are not independent corroboration.", "source": "https://arxiv.org/html/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Rewrote the short title and summary in plain language, preserving the reported benefit and material limits. Added the named publication or institution to the reviewed lead source for homepage display.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Identified the AI system or feature in the supporting summary, or stated that the source does not name the software. Preserved the reported benefit and material qualifications.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Shortened the homepage supporting statement to one sentence naming the system and its contribution; retained detailed outcomes and qualifications in the case account and source notes.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Revised the title and subtitle for natural phrasing, with named researchers, shorter dependencies, and less compressed terminology; preserved the system identities and evidence qualifications.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Identified MADGRAV-2D and recorded the authors' final pipeline configuration dated 14 September 2026. Removed the paper revision from the AI model-version field. Added a fixed repository snapshot, checkpoint identifier, and calibration environment to the configuration metadata; preserved the arXiv versions as publication metadata. No separately labeled model release or exact runtime commit is established.", "source": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Revised the homepage subtitle for concise database prose, with courtesy titles for named beneficiaries and numbered clauses for the Coates drug-selection account. Preserved the reported evidence and material uncertainty.", "source": null, "editor": "AIBD" }, { "date": "2026-10-25", "change": "Added both Austrian and Italian developer affiliations and identified Austria as the documented research-computing location. Distinguished the U.S. LIGO data sources from the research location. Preserved the first-version publication date and counted the later paper version as the same source.", "source": "https://arxiv.org/html/2609.39583v2", "editor": "AIBD" } ], "confidence": "claimed", "shade": "low", "claim_status": "The makers' claim", "claim_basis": "Sources: the makers' preprint and the makers' other. The cited evidence includes no independent source and no finding by a regulator or court.", "profile": { "case_id": "AIBD-0283", "catalog_version": "1.0.0", "source_fingerprint": "9a239a727d550cd771e1f73eca2daaf960cfa4c0c3a9881269f88fa1e5a365a0", "collection_refs": [ { "collection_id": "aibd", "version": "snapshot-8bf7cfec9cb185aa" }, { "collection_id": "aibd", "version": "2026-10-25" } ], "source_ids": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "families": { "strategic assessments": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "record_ids": [] }, "evaluations": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "record_ids": [] }, "monetary observations": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "record_ids": [] }, "institutional decisions": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "record_ids": [] } }, "fields": { "actors[]": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "developer": "Marietta Blau Institute for Particle Physics, Austrian Academy of Sciences, with the University of Urbino and INFN Florence", "deployer": "The same authors, on public LIGO O3 and O4 data" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "developer", "deployer" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "actors[].ownership_control": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "ai_contribution": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ai_role": { "status": "assessed", "statement": "The convolutional networks performed the whole detection chain, from anomaly detection to ranking, and the authors are the developers. Every detection matches a published catalog event. The authors state that no candidate without a catalog counterpart survives, so the search added no new discovery in this analysis." } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "ai_role" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "ai_relationship": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "category": "general benefit", "ai_role": { "status": "assessed", "statement": "The convolutional networks performed the whole detection chain, from anomaly detection to ranking, and the authors are the developers. Every detection matches a published catalog event. The authors state that no candidate without a catalog counterpart survives, so the search added no new discovery in this analysis." } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "ai_role", "category" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "assessments[].alternatives": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].assessor_date": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].depth": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].display_confidence": { "state": "value_present", "value": { "kind": "structured", "data": { "confidence": "low", "evidence_word": "claimed", "claim_status": "The makers' claim", "basis": "Sources: the makers' preprint and the makers' other. The cited evidence includes no independent source and no finding by a regulator or court.", "interpretation": "The display category follows the existing rule. Source lineage and claim-specific limitations are stated here; a publication count is not a count of independent confirmations." } }, "provenance": { "kind": "source_review", "basis": "Rechecked source origin and independence. The display classification remains governed by the existing record rule.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "partial" }, "assessments[].rationale": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ratings": [ { "group": "organizational", "reason": "Recovery of 47 known events, 44 of them shared with coherent WaveBurst, demonstrates the method. No new event, catalog adoption or operational use is documented." }, { "group": "societal", "reason": "A complementary search channel is proposed; no scientific result beyond re-detection is claimed." } ] } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "ratings" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "assessments[].rubric_version": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].status_basis": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "assessments[].target": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ratings": [ { "group": "organizational", "reason": "Recovery of 47 known events, 44 of them shared with coherent WaveBurst, demonstrates the method. No new event, catalog adoption or operational use is documented." }, { "group": "societal", "reason": "A complementary search channel is proposed; no scientific result beyond re-detection is claimed." } ] } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "ratings" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "availability": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "brief": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "case_kind": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "case_links[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "claims[].evidence_links": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "claims[].evidence_strength": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "claims[].missingness": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "claims[].proposition": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "cross_border_flows[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "decisions[].action_outcome": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "institutional decisions" }, "decisions[].grounds": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "institutional decisions" }, "decisions[].institution_task": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "institutional decisions" }, "decisions[].success_criterion": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "institutional decisions" }, "deployment_stage": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "economics[].amount": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].costs[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].counterfactual": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].currency_price_basis": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].gross_net": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].measure_type": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].period_unit": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].realization": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].recipient_payer": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].return_payback": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "economics[].valuation_method": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "monetary observations" }, "evaluations[].ai_uplift": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].comparator": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].design": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].effect": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].exposure": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].id": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].measurement_method": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].metric": { "state": "not_assessed", "provenance": { "kind": "legacy_record", "basis": "Mapped from the existing case description without a new external source review.", "record_fields": [ "metric" ], "source_refs": [] }, "review_state": "not_assessed", "coverage": "partial", "family": "evaluations", "display": { "answer": "47 detections with a false alarm rate below one per year and p_astro above 0.9", "editor_note": "" } }, "evaluations[].reproduction_materials": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].setting_sample": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "evaluations[].uncertainty_limitations": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "evaluations" }, "external_references[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "follow_up[]": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "later": { "status": "unknown", "note": "The authors revised the preprint on 19 October 2026. No independent follow-up result is cited." } } }, "provenance": { "kind": "source_review", "basis": "Updated the affected source references, dates, or account after reading the original sources. Other mapped content and its review status remain unchanged.", "record_fields": [ "later" ], "source_refs": [ "AIBD-0283-source-0001" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "partial" }, "human_oversight": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "id": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "id": "AIBD-0283" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "id" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "infrastructure_dependencies[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].beneficiaries": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].beneficiary_count": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].benefit_type": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "benefit_type": "Improved Task Accuracy", "taxonomy_note": "Re-detection of cataloged events is read as task performance; no discovery is claimed." } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "benefit_type", "taxonomy_note" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "outcomes[].claim": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "account": "A paper was posted on arXiv on 30 September 2026 by Gianluca Inguglia, Huw Haigh and Ulyana Dupletsa. They are at the Marietta Blau Institute for Particle Physics (Austrian Academy of Sciences, Vienna). Alessandro Longo (University of Urbino and INFN Florence) is a co-author. It is titled “MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data” (version 2, 19 October 2026; submitted to the European Physical Journal C). The pipeline is “a series of sequential convolutional neural networks that perform anomaly detection, glitch classification, coherence testing, and signal ranking”. It is applied to one-second Q-transform windows of Hanford and Livingston data from O3a, O3b, O4a and O4b. That is 442.9 days of coincident livetime in total. The abstract reports 47 detections with a false alarm rate below one per year and p_astro above 0.9. Of those, “44 are shared with the minimally modeled coherent WaveBurst search”. The three others are cataloged events, and the paper states that “no candidate without a catalog counterpart survives the analysis”. Recovery rises with mass. It reaches 8.1% of confident catalog events below 30 solar masses and 39.8% between 30 and 100. It reaches 53.3% above 100 for network SNR above 10. The authors conclude that anomaly-detection pipelines “can serve as an independent detection channel complementary to matched filtering in the high-mass high-SNR regime”. The paper confirms no new event independently. The authors identify the two-detector configuration as MADGRAV-2D in their code repository. Its changelog dates the final pipeline configuration to 14 September 2026, and its README states that this configuration produced the reported 47 detections. The repository provides the trained weights and frozen calibration.", "benefit_type": "Improved Task Accuracy" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "account", "benefit_type" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "outcomes[].distribution": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].duration": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].id": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].stakeholder_group": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "ratings": [ { "group": "organizational", "reason": "Recovery of 47 known events, 44 of them shared with coherent WaveBurst, demonstrates the method. No new event, catalog adoption or operational use is documented." }, { "group": "societal", "reason": "A complementary search channel is proposed; no scientific result beyond re-detection is claimed." } ] } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "ratings" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "outcomes[].status": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].tradeoffs": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "outcomes[].unexpected": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "unexpected": false, "purpose": "MADGRAV was built to search LIGO data for high-mass compact binary coalescences by anomaly detection. Recovering cataloged events was that purpose." } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "unexpected", "purpose" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "places[]": { "state": "value_present", "value": { "kind": "structured", "data": { "location": "Vienna, Austria", "developer_country": "Austria", "deployer_country": "Austria", "developer_jurisdictions": [ "Austria", "Italy" ], "case_jurisdictions": [ "Austria" ], "location_label": "Research location", "location_note": "The paper identifies Austrian computing facilities. The LIGO instruments supplied U.S. data." } }, "provenance": { "kind": "source_review", "basis": "The original paper’s author affiliations identify the Marietta Blau Institute in Vienna, Austria, and the University of Urbino and INFN Florence in Italy. The acknowledgments state that computational results used Austrian Scientific Computing infrastructure, including VSC-5 and MUSICA. Austria is the documented research-computing location, not the location of the celestial events. Hanford and Livingston LIGO supplied U.S. observational data. This review adds geographic roles from version 2 without treating it as an independent work or changing the version 1 publication chronology.", "record_fields": [ "location", "developer_country", "deployer_country" ], "source_refs": [ "AIBD-0283-source-0001" ], "reviewed_on": "2026-10-25" }, "review_state": "reviewed", "coverage": "partial", "display": { "answer": "The research team had Austrian and Italian affiliations and used computing facilities in Austria. LIGO instruments in the United States supplied the observations.", "editor_note": "" } }, "purpose": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "purpose": "MADGRAV was built to search LIGO data for high-mass compact binary coalescences by anomaly detection. Recovering cataloged events was that purpose." } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "purpose" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "record_dates": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "date_observed": "2026-09-30", "date_basis": "source publication", "date_precision": "day", "revisions": [ { "date": "2026-10-20", "change": "Record drafted from the cited sources.", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Title reworded to state the event without naming who claims it.", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Added what the benefit is measured against from the cited sources.", "source": "https://arxiv.org/html/2609.39583v2", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Benefit type set to Improved Task Accuracy under Trustible's full taxonomy of 23 types; it was Improved Knowledge Discovery among the six first used.", "source": "https://trustible.ai/post/a-framework-for-measuring-the-benefits-of-ai/", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the developer and the deployer added from the organization's own site or a registry.", "source": "https://inspirehep.net/api/institutions?q=Marietta%20Blau&size=3; https://inspirehep.net/api/institutions?q=Marietta%20Blau&size=3", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Prose pass: a rating reason, account, purpose, the system's part rewritten in shorter sentences with abbreviations expanded; facts, figures and quotations unchanged.", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Notes about which sources could be read were reworded for readers; facts, figures and quotations unchanged.", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Consolidated the abstract and HTML citations into one primary preprint. Read version 1 and verified that the 47 detections, 44 shared detections, 442.9 days of data, recovery rates, comparison percentages, and absence of a new candidate were already reported on 30 September 2026. Preserved the 19 October 2026 version 2 link and the submission history in source provenance. Versions of the same work are not independent corroboration.", "source": "https://arxiv.org/html/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Rewrote the short title and summary in plain language, preserving the reported benefit and material limits. Added the named publication or institution to the reviewed lead source for homepage display.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Identified the AI system or feature in the supporting summary, or stated that the source does not name the software. Preserved the reported benefit and material qualifications.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Shortened the homepage supporting statement to one sentence naming the system and its contribution; retained detailed outcomes and qualifications in the case account and source notes.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Revised the title and subtitle for natural phrasing, with named researchers, shorter dependencies, and less compressed terminology; preserved the system identities and evidence qualifications.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Identified MADGRAV-2D and recorded the authors' final pipeline configuration dated 14 September 2026. Removed the paper revision from the AI model-version field. Added a fixed repository snapshot, checkpoint identifier, and calibration environment to the configuration metadata; preserved the arXiv versions as publication metadata. No separately labeled model release or exact runtime commit is established.", "source": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Revised the homepage subtitle for concise database prose, with courtesy titles for named beneficiaries and numbered clauses for the Coates drug-selection account. Preserved the reported evidence and material uncertainty.", "editor": "AIBD" }, { "date": "2026-10-25", "change": "Added both Austrian and Italian developer affiliations and identified Austria as the documented research-computing location. Distinguished the U.S. LIGO data sources from the research location. Preserved the first-version publication date and counted the later paper version as the same source.", "source": "https://arxiv.org/html/2609.39583v2", "editor": "AIBD" } ] } }, "provenance": { "kind": "source_review", "basis": "Added the geography review to the revision history and preserved the original publication chronology and prior version assessment.", "record_fields": [ "date_observed", "revisions", "date_basis", "date_precision" ], "source_refs": [ "AIBD-0283-source-0001" ], "reviewed_on": "2026-10-25" }, "review_state": "reviewed", "coverage": "partial" }, "replication.attempts[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "replication.conditions": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "replication": { "produced_by": [ "Sequential CNNs for anomaly detection, glitch classification, coherence testing and signal ranking on one-second Q-transform windows of Hanford and Livingston data", "442.9 days of coincident livetime across O3a, O3b, O4a and O4b" ], "must_hold": [ "High-mass, high-SNR sources: the recovered fraction rises from 8.1% below 30 solar masses to 53.3% above 100", "Two-detector coherence between Hanford and Livingston" ], "echoes": [] } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "replication" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "revision_history[]": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "revisions": [ { "date": "2026-10-20", "change": "Record drafted from the cited sources.", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Title reworded to state the event without naming who claims it.", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Added what the benefit is measured against from the cited sources.", "source": "https://arxiv.org/html/2609.39583v2", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Benefit type set to Improved Task Accuracy under Trustible's full taxonomy of 23 types; it was Improved Knowledge Discovery among the six first used.", "source": "https://trustible.ai/post/a-framework-for-measuring-the-benefits-of-ai/", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Country of the developer and the deployer added from the organization's own site or a registry.", "source": "https://inspirehep.net/api/institutions?q=Marietta%20Blau&size=3; https://inspirehep.net/api/institutions?q=Marietta%20Blau&size=3", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Prose pass: a rating reason, account, purpose, the system's part rewritten in shorter sentences with abbreviations expanded; facts, figures and quotations unchanged.", "editor": "AIBD" }, { "date": "2026-10-21", "change": "Notes about which sources could be read were reworded for readers; facts, figures and quotations unchanged.", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Consolidated the abstract and HTML citations into one primary preprint. Read version 1 and verified that the 47 detections, 44 shared detections, 442.9 days of data, recovery rates, comparison percentages, and absence of a new candidate were already reported on 30 September 2026. Preserved the 19 October 2026 version 2 link and the submission history in source provenance. Versions of the same work are not independent corroboration.", "source": "https://arxiv.org/html/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Rewrote the short title and summary in plain language, preserving the reported benefit and material limits. Added the named publication or institution to the reviewed lead source for homepage display.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Identified the AI system or feature in the supporting summary, or stated that the source does not name the software. Preserved the reported benefit and material qualifications.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Shortened the homepage supporting statement to one sentence naming the system and its contribution; retained detailed outcomes and qualifications in the case account and source notes.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Revised the title and subtitle for natural phrasing, with named researchers, shorter dependencies, and less compressed terminology; preserved the system identities and evidence qualifications.", "source": "https://arxiv.org/abs/2609.39583v1", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Identified MADGRAV-2D and recorded the authors' final pipeline configuration dated 14 September 2026. Removed the paper revision from the AI model-version field. Added a fixed repository snapshot, checkpoint identifier, and calibration environment to the configuration metadata; preserved the arXiv versions as publication metadata. No separately labeled model release or exact runtime commit is established.", "source": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "editor": "AIBD" }, { "date": "2026-10-24", "change": "Revised the homepage subtitle for concise database prose, with courtesy titles for named beneficiaries and numbered clauses for the Coates drug-selection account. Preserved the reported evidence and material uncertainty.", "editor": "AIBD" }, { "date": "2026-10-25", "change": "Added both Austrian and Italian developer affiliations and identified Austria as the documented research-computing location. Distinguished the U.S. LIGO data sources from the research location. Preserved the first-version publication date and counted the later paper version as the same source.", "source": "https://arxiv.org/html/2609.39583v2", "editor": "AIBD" } ] } }, "provenance": { "kind": "source_review", "basis": "Added the geography review to the revision history and preserved the original publication chronology and prior version assessment.", "record_fields": [ "revisions" ], "source_refs": [ "AIBD-0283-source-0001" ], "reviewed_on": "2026-10-25" }, "review_state": "reviewed", "coverage": "partial" }, "sector": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "domain": "Astrophysics / gravitational-wave astronomy", "segment": "Academic research on public detector data" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "domain", "segment" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "sources[].access_version": { "state": "value_present", "value": { "kind": "structured", "data": { "source_ref": "AIBD-0283-source-0001", "accessed_on": "2026-10-24", "first_submission": { "version": "2609.39583v1", "submitted_at": "2026-09-30T12:12:19Z", "url": "https://arxiv.org/html/2609.39583v1" }, "latest_revision_read": { "version": "2609.39583v2", "submitted_at": "2026-10-19T10:02:39Z", "url": "https://arxiv.org/html/2609.39583v2" }, "first_version_support": [ { "locator": "Abstract", "supports": "47 detections, 44 shared with coherent WaveBurst, and mass-dependent recovery rates." }, { "locator": "Section 3", "supports": "442.9 days of coincident detector data." }, { "locator": "Section 4", "supports": "No candidate without a catalog counterpart survives the analysis." }, { "locator": "Section 5", "supports": "17 percent of the catalog population and 39 percent of the coherent WaveBurst count." } ], "date_basis": "The case date is the first publication of the reported findings, verified in version 1; the revision date is recorded separately.", "software_source": { "source_ref": "AIBD-0283-source-0002", "url": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "commit": "acb66841f9408939c1ca17bff90360214c4ea2c9", "commit_date": "2026-09-19", "accessed_on": "2026-10-24", "qualification": "A fixed code snapshot, separate from the preprint versions; the exact runtime commit is not established." }, "reviewed_on": "2026-10-25", "versions_and_access": [ { "source_ref": "AIBD-0283-source-0001", "source_url": "https://arxiv.org/abs/2609.39583v1", "access_url": "https://arxiv.org/html/2609.39583v2", "publication_date": "2026-10-19", "accessed_on": "2026-10-25", "locator": "Author affiliations and Acknowledgements, paragraph 1.", "status": "Version 2 inspected for developer affiliations and documented computing facilities. Same work as the first version, not another source." } ] } }, "provenance": { "kind": "source_review", "basis": "Inspected the geography passages in version 2 and retained the existing first-version chronology.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-25" }, "review_state": "reviewed", "coverage": "complete" }, "sources[].citation_locator": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "evidence": [ { "kind": "preprint", "citation": "Inguglia G, Haigh H, Dupletsa U, Longo A (2026). MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data. arXiv:2609.39583v1, 30 September 2026, abstract and sections 3–5. Revised as v2 on 19 October 2026. Authored by the pipeline's developers.", "title": "MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data", "url": "https://arxiv.org/abs/2609.39583v1", "date": "2026-09-30", "written_by": "makers", "quote": "Of the 47 detections, 44 are shared with the minimally modelled coherent WaveBurst search." }, { "kind": "other", "citation": "Inguglia, Gianluca, and collaborators. MADGRAV code repository, snapshot acb66841f9408939c1ca17bff90360214c4ea2c9, 19 September 2026. README opening and sections 3–6; CHANGELOG entry for 14 September 2026; spectrogram_cascade/massive_calibration_BA.json; environment.yml.", "title": "MADGRAV: Multilevel Anomaly Detection for GRAVitational-wave science", "url": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "date": "2026-09-19", "written_by": "makers", "quote": "This is the final version of the pipeline" } ] } }, "provenance": { "kind": "source_review", "basis": "Updated the affected source references, dates, or account after reading the original sources. Other mapped content and its review status remain unchanged.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0283-source-0001" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "partial" }, "sources[].id": { "state": "value_present", "value": { "kind": "identifiers", "identifiers": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "role": "Active evidence sources in record order" }, "provenance": { "kind": "curatorial", "basis": "Preserved existing source identifiers and appended the material technical source where applicable.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "complete" }, "sources[].language": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "sources[].relationship": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "evidence": [ { "kind": "preprint", "citation": "Inguglia G, Haigh H, Dupletsa U, Longo A (2026). MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data. arXiv:2609.39583v1, 30 September 2026, abstract and sections 3–5. Revised as v2 on 19 October 2026. Authored by the pipeline's developers.", "title": "MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data", "url": "https://arxiv.org/abs/2609.39583v1", "date": "2026-09-30", "written_by": "makers", "quote": "Of the 47 detections, 44 are shared with the minimally modelled coherent WaveBurst search." }, { "kind": "other", "citation": "Inguglia, Gianluca, and collaborators. MADGRAV code repository, snapshot acb66841f9408939c1ca17bff90360214c4ea2c9, 19 September 2026. README opening and sections 3–6; CHANGELOG entry for 14 September 2026; spectrogram_cascade/massive_calibration_BA.json; environment.yml.", "title": "MADGRAV: Multilevel Anomaly Detection for GRAVitational-wave science", "url": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "date": "2026-09-19", "written_by": "makers", "quote": "This is the final version of the pipeline" } ] } }, "provenance": { "kind": "source_review", "basis": "Updated the affected source references, dates, or account after reading the original sources. Other mapped content and its review status remain unchanged.", "record_fields": [ "evidence" ], "source_refs": [ "AIBD-0283-source-0001" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "partial" }, "sources[].upstream_sources": { "state": "value_present", "value": { "kind": "structured", "data": { "origins": [ { "source_refs": [ "AIBD-0283-source-0001" ], "url": "https://arxiv.org/abs/2609.39583v1", "publisher": "arXiv", "authors": [ "Gianluca Inguglia", "Huw Haigh", "Ulyana Dupletsa", "Alessandro Longo" ], "published_on": "2026-09-30", "relationship": "One primary preprint by the pipeline developers. Its abstract, HTML versions, and revision are the same work, not independent evidence.", "display_name": "arXiv", "contribution": " Version 2 author affiliations establish Austria and Italy, and acknowledgments identify Austrian computing facilities.", "locator": " Author affiliations and Acknowledgements, paragraph 1." }, { "source_refs": [ "AIBD-0283-source-0002" ], "url": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "publisher": "Gianluca Inguglia and the MADGRAV research team", "published_on": "2026-09-19", "relationship": "Original code, trained weights, and frozen configuration linked from section 2 of the paper. The repository identifies MADGRAV-2D and the final configuration that produced the reported detections. This supplies software provenance, not independent corroboration.", "display_name": "MADGRAV research team" } ], "versions": [ { "version": "2609.39583v1", "published_on": "2026-09-30", "url": "https://arxiv.org/html/2609.39583v1" }, { "version": "2609.39583v2", "published_on": "2026-10-19", "url": "https://arxiv.org/html/2609.39583v2" } ], "independent_evaluations_of_reported_results": 0, "lead_source_ref": "AIBD-0283-source-0001", "retained_source_order": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "ordering_basis": "The first retained source is the original publication that supports the main reported benefit. Any additional retained source follows it and supplies the separate evidence described in its relationship.", "software_provenance": { "source_ref": "AIBD-0283-source-0002", "paper_locator": "Section 2, citation to Inguglia et al. (2026a)", "snapshot_commit": "acb66841f9408939c1ca17bff90360214c4ea2c9", "configuration_date": "2026-09-14", "files": [ { "url": "https://github.com/ginguglia/MADGRAV/blob/acb66841f9408939c1ca17bff90360214c4ea2c9/README.md", "locator": "Opening paragraphs, Important note, and sections 3–6", "supports": "MADGRAV-2D identity and the connection between the final configuration and the reported 47 detections." }, { "url": "https://github.com/ginguglia/MADGRAV/blob/acb66841f9408939c1ca17bff90360214c4ea2c9/CHANGELOG.md", "locator": "Entry for 14 September 2026", "supports": "The dated final configuration, including the injection-set glitch-arm gate." }, { "url": "https://github.com/ginguglia/MADGRAV/blob/acb66841f9408939c1ca17bff90360214c4ea2c9/spectrogram_cascade/massive_calibration_BA.json", "locator": "model_path", "supports": "The deployed autoencoder checkpoint path." }, { "url": "https://github.com/ginguglia/MADGRAV/blob/acb66841f9408939c1ca17bff90360214c4ea2c9/environment.yml", "locator": "Frozen-weights training/calibration comments", "supports": "PyTorch, CUDA, and cuDNN versions for calibration of the frozen weights." } ], "limitations": [ "The authors do not identify the exact runtime commit.", "The code repository and paper have the same authorship and are not independent confirmations.", "There is no numbered software release or Git tag." ] }, "reviewed_on": "2026-10-25", "claim_chains": [ { "claim": "Developer affiliations and research location", "path": [ "https://arxiv.org/html/2609.39583v2" ], "finding": "The original paper’s author affiliations identify the Marietta Blau Institute in Vienna, Austria, and the University of Urbino and INFN Florence in Italy. The acknowledgments state that computational results used Austrian Scientific Computing infrastructure, including VSC-5 and MUSICA. Austria is the documented research-computing location, not the location of the celestial events. Hanford and Livingston LIGO supplied U.S. observational data. This review adds geographic roles from version 2 without treating it as an independent work or changing the version 1 publication chronology." } ], "review_scope": " Additional geographic review on October 25, 2026, distinguishes author affiliations, research computing, and the observational-data source without adding another independent work." } }, "provenance": { "kind": "source_review", "basis": "Added geographic roles from the original paper’s affiliations and acknowledgments, preserving the existing version chronology and source identity.", "record_fields": [ "evidence", "developer_country", "location" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-25" }, "review_state": "reviewed", "coverage": "complete", "display": { "answer": "The developers’ preprint reports the detections. Their repository supplies the code, trained weights, and configuration. The paper’s revisions and code are supporting material from the same team.", "editor_note": "" } }, "strategic_assessments[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial", "family": "strategic assessments" }, "summary": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "summary": "Researchers used MADGRAV-2D to recover 47 cataloged gravitational-wave signals from LIGO data.", "account": "A paper was posted on arXiv on 30 September 2026 by Gianluca Inguglia, Huw Haigh and Ulyana Dupletsa. They are at the Marietta Blau Institute for Particle Physics (Austrian Academy of Sciences, Vienna). Alessandro Longo (University of Urbino and INFN Florence) is a co-author. It is titled “MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data” (version 2, 19 October 2026; submitted to the European Physical Journal C). The pipeline is “a series of sequential convolutional neural networks that perform anomaly detection, glitch classification, coherence testing, and signal ranking”. It is applied to one-second Q-transform windows of Hanford and Livingston data from O3a, O3b, O4a and O4b. That is 442.9 days of coincident livetime in total. The abstract reports 47 detections with a false alarm rate below one per year and p_astro above 0.9. Of those, “44 are shared with the minimally modeled coherent WaveBurst search”. The three others are cataloged events, and the paper states that “no candidate without a catalog counterpart survives the analysis”. Recovery rises with mass. It reaches 8.1% of confident catalog events below 30 solar masses and 39.8% between 30 and 100. It reaches 53.3% above 100 for network SNR above 10. The authors conclude that anomaly-detection pipelines “can serve as an independent detection channel complementary to matched filtering in the high-mass high-SNR regime”. The paper confirms no new event independently. The authors identify the two-detector configuration as MADGRAV-2D in their code repository. Its changelog dates the final pipeline configuration to 14 September 2026, and its README states that this configuration produced the reported 47 detections. The repository provides the trained weights and frozen calibration." } }, "provenance": { "kind": "curatorial", "basis": "Updated the displayed wording and revision history without changing the reviewed facts, sources, or assessments.", "record_fields": [ "summary", "account" ], "source_refs": [], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "partial" }, "systems[].autonomy_permissions": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "systems[].component_role": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "systems[].deployment_configuration": { "state": "value_present", "value": { "kind": "structured", "data": { "pipeline": "MADGRAV-2D", "detectors": [ "LIGO Hanford", "LIGO Livingston" ], "observing_runs": [ "O3a", "O3b", "O4a", "O4b" ], "configuration_date": "2026-09-14", "configuration_label": "Final version: injection-set glitch-arm gate, consistent injection-side background", "configuration_basis": { "source_ref": "AIBD-0283-source-0002", "locators": [ "README opening and sections 3–6", "CHANGELOG entry for 14 September 2026" ] }, "code_snapshot": { "repository": "https://github.com/ginguglia/MADGRAV", "commit": "acb66841f9408939c1ca17bff90360214c4ea2c9", "commit_date": "2026-09-19", "url": "https://github.com/ginguglia/MADGRAV/tree/acb66841f9408939c1ca17bff90360214c4ea2c9", "role": "Reviewed public snapshot of the authors' code, weights, and configuration", "exact_runtime_commit_established": false }, "final_configuration_commit": "89cf283954bc865515d31f28ae604976517776f4", "model_components": [ { "implementation_name": "BaselineCAE", "architecture": "Convolutional autoencoder", "checkpoint": "assets/models/baseline_cae_weaksup_best.pt", "checkpoint_git_blob_sha1": "ddb9e78927a6fc3b1bea139d9b6f548f5065f3f3", "configuration": "spectrogram_cascade/massive_calibration_BA.json", "configuration_key": "model_path", "source_ref": "AIBD-0283-source-0002" }, { "role": "Glitch classification", "architecture": "Five-seed convolutional-neural-network ensemble", "checkpoints": [ "lr_cascade/p1v42/arm_deploy_seed0.pt", "lr_cascade/p1v42/arm_deploy_seed1.pt", "lr_cascade/p1v42/arm_deploy_seed2.pt", "lr_cascade/p1v42/arm_deploy_seed3.pt", "lr_cascade/p1v42/arm_deploy_seed4.pt" ], "source_ref": "AIBD-0283-source-0002" }, { "role": "Low- and high-frequency specialist classification", "architecture": "Two convolutional neural networks", "source_ref": "AIBD-0283-source-0001", "locator": "Section 2.1, Architecture and training" } ], "training_and_calibration_environment": { "pytorch": "1.12.1.post201", "cuda": "11.2", "cudnn": "8.9.0.7", "source_ref": "AIBD-0283-source-0002", "locator": "environment.yml, frozen-weights training/calibration comment" }, "qualification": "The authors link this configuration to the reported results, but do not specify the exact runtime commit. The Git blob identifies the checkpoint file; it is not a model release number. The arXiv revisions belong to the paper." } }, "provenance": { "kind": "source_review", "basis": "Traced the paper's repository citation to the authors' README, dated changelog, frozen model configuration, checkpoint listing, and calibration environment.", "record_fields": [ "system", "model_version" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "complete", "display": { "answer": "MADGRAV-2D combined anomaly detection with classifiers for glitches and low- and high-frequency signals. It analyzed observations from the Hanford and Livingston detectors during LIGO’s O3 and O4 runs.", "editor_note": "" } }, "systems[].inputs_outputs": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "systems[].model_name": { "state": "value_present", "value": { "kind": "structured", "data": { "named_component": "BaselineCAE", "component_role": "Convolutional autoencoder for anomaly detection in the MADGRAV-2D cascade.", "scope": "An implementation name in the authors’ published code. The full cascade includes other convolutional networks, and this name does not identify a release." } }, "provenance": { "kind": "source_review", "basis": "The authors’ configuration and code identify BaselineCAE and its checkpoint as a component of the published MADGRAV-2D pipeline.", "record_fields": [ "system" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "partial", "display": { "answer": "BaselineCAE, a convolutional autoencoder, together with convolutional networks that classify glitches and signals.", "editor_note": "" } }, "systems[].model_version": { "state": "not_reported", "provenance": { "kind": "source_review", "basis": "Read the paper and the authors' repository, including its README, changelog, frozen calibration configuration, checkpoint listing, and release history.", "record_fields": [ "model_version" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-24" }, "reason": "The authors publish trained weights and a dated pipeline configuration, but no separately labeled model release for the reported analysis. The configuration date, checkpoint file, and reviewed repository snapshot are recorded separately; none is substituted for an undisclosed model release.", "review_state": "reviewed", "coverage": "complete", "display": { "answer": "The authors publish trained weights and a configuration dated 14 September 2026, without a separate model release number.", "editor_note": "The dated configuration comes from the authors’ repository changelog (reference 2). Their published checkpoint and code snapshot identify available files but do not establish the exact code commit used for every reported run." } }, "systems[].product_name": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "system": "MADGRAV-2D (two-detector cascade of convolutional neural networks)" } }, "provenance": { "kind": "source_review", "basis": "Identified the reported product or pipeline from its original case and technical sources; kept hardware, underlying model, and software release distinct.", "record_fields": [ "system" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "complete", "display": { "answer": "MADGRAV-2D", "editor_note": "" } }, "systems[].provider": { "state": "value_present", "value": { "kind": "text", "text": "Gianluca Inguglia and the MADGRAV research team" }, "provenance": { "kind": "source_review", "basis": "The developers authored the paper and provide the public code and trained weights through Gianluca Inguglia's repository.", "record_fields": [ "developer" ], "source_refs": [ "AIBD-0283-source-0001", "AIBD-0283-source-0002" ], "reviewed_on": "2026-10-24" }, "review_state": "reviewed", "coverage": "complete" }, "taxonomy_crosswalks[]": { "state": "not_assessed", "review_state": "not_assessed", "coverage": "partial" }, "title": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "title": "MADGRAV anomaly-detection pipeline recovers 47 gravitational-wave detections in LIGO O3 and O4 data, 44 shared with coherent WaveBurst and none new", "headline": "Researchers found 47 known gravitational-wave signals" } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "title", "headline" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" }, "version": { "state": "value_present", "value": { "kind": "identifiers", "identifiers": [ "9a239a727d550cd771e1f73eca2daaf960cfa4c0c3a9881269f88fa1e5a365a0" ], "role": "SHA-256 of normalized original case data" }, "provenance": { "kind": "curatorial", "basis": "Recomputed the fingerprint after the homepage copy revision.", "record_fields": [ "summary", "revisions" ], "source_refs": [], "reviewed_on": "2026-10-24" }, "review_state": "record_mapped", "coverage": "complete" }, "workflow": { "state": "value_present", "value": { "kind": "record_fields", "fields": { "account": "A paper was posted on arXiv on 30 September 2026 by Gianluca Inguglia, Huw Haigh and Ulyana Dupletsa. They are at the Marietta Blau Institute for Particle Physics (Austrian Academy of Sciences, Vienna). Alessandro Longo (University of Urbino and INFN Florence) is a co-author. It is titled “MADGRAV: a multilevel anomaly-detection pipeline for gravitational-wave searches applied to LIGO data” (version 2, 19 October 2026; submitted to the European Physical Journal C). The pipeline is “a series of sequential convolutional neural networks that perform anomaly detection, glitch classification, coherence testing, and signal ranking”. It is applied to one-second Q-transform windows of Hanford and Livingston data from O3a, O3b, O4a and O4b. That is 442.9 days of coincident livetime in total. The abstract reports 47 detections with a false alarm rate below one per year and p_astro above 0.9. Of those, “44 are shared with the minimally modeled coherent WaveBurst search”. The three others are cataloged events, and the paper states that “no candidate without a catalog counterpart survives the analysis”. Recovery rises with mass. It reaches 8.1% of confident catalog events below 30 solar masses and 39.8% between 30 and 100. It reaches 53.3% above 100 for network SNR above 10. The authors conclude that anomaly-detection pipelines “can serve as an independent detection channel complementary to matched filtering in the high-mass high-SNR regime”. The paper confirms no new event independently. The authors identify the two-detector configuration as MADGRAV-2D in their code repository. Its changelog dates the final pipeline configuration to 14 September 2026, and its README states that this configuration produced the reported 47 detections. The repository provides the trained weights and frozen calibration.", "replication": { "produced_by": [ "Sequential CNNs for anomaly detection, glitch classification, coherence testing and signal ranking on one-second Q-transform windows of Hanford and Livingston data", "442.9 days of coincident livetime across O3a, O3b, O4a and O4b" ], "must_hold": [ "High-mass, high-SNR sources: the recovered fraction rises from 8.1% below 30 solar masses to 53.3% above 100", "Two-detector coherence between Hanford and Livingston" ], "echoes": [] } } }, "provenance": { "kind": "legacy_record", "basis": "Preserved from the existing AIBD record; external sources were not rechecked.", "record_fields": [ "account", "replication" ], "source_refs": [] }, "review_state": "record_mapped", "coverage": "partial" } }, "summary": { "total": 91, "addressed": 91, "known": 31, "reviewed": 15, "record_mapped": 17, "not_assessed": 59, "complete": false, "by_state": { "value_present": 31, "not_reported": 1, "unresolved": 0, "not_assessed": 59, "not_applicable": 0, "withheld": 0, "source_unavailable": 0 } } } }