{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SNWUXK5XLHWKECHTP6XQ7RCCE7","short_pith_number":"pith:SNWUXK5X","schema_version":"1.0","canonical_sha256":"936d4babb759eca208f37faf0fc44227c8fe885a1b37182d1779789c266bff9b","source":{"kind":"arxiv","id":"2605.02651","version":2},"attestation_state":"computed","paper":{"title":"ARA: Agentic Reproducibility Assessment For Scalable Support Of Scientific Peer-Review","license":"http://creativecommons.org/licenses/by/4.0/","headline":"ARA extracts directed workflow graphs from papers to evaluate reproducibility at scale.","cross_cats":["cs.LG"],"primary_cat":"cs.DL","authors_text":"Anastasios Kouvelas, Andres L. Marin, Fan Wu, Georgios Fontaras, Kevin Riehl, Michail A. Makridis, Nikofors Zacharof, Patrick Langer, Robert Jakob","submitted_at":"2026-05-04T14:34:36Z","abstract_excerpt":"Scientific peer review increasingly struggles to assess reproducibility at the scale and complexity of modern research output. Evaluating reproducibility requires reconstructing experimental dependencies, methodological choices, data flows, and result-generating procedures, which often exceeds what human reviewers can provide. Agentic Reproducibility Assessment (ARA) formalizes reproducibility assessment as a structured reasoning task over scientific documents. Given a paper, ARA extracts a directed workflow graph linking sources, methods, experiments, and outputs, then evaluates its reconstru"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.02651","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.DL","submitted_at":"2026-05-04T14:34:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"1530913cd0ae2ca7c7b7034cd6d1daeb1e2835f61cbfff88c2190e62d7dafae0","abstract_canon_sha256":"c9d6d8e7d866c0884e17b18808aaac4ba6367dadc6f3e6da904104a9e3f316c1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:00:40.406215Z","signature_b64":"683t02mLRxEvfsStcqGwrWtXpeO0spPtVqH+riG67HuSoN7DVm6fiKPT+jS0Kbt3i2Q6fs4TcQyfMVw/b68rBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"936d4babb759eca208f37faf0fc44227c8fe885a1b37182d1779789c266bff9b","last_reissued_at":"2026-05-20T00:00:40.405373Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:00:40.405373Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ARA: Agentic Reproducibility Assessment For Scalable Support Of Scientific Peer-Review","license":"http://creativecommons.org/licenses/by/4.0/","headline":"ARA extracts directed workflow graphs from papers to evaluate reproducibility at scale.","cross_cats":["cs.LG"],"primary_cat":"cs.DL","authors_text":"Anastasios Kouvelas, Andres L. Marin, Fan Wu, Georgios Fontaras, Kevin Riehl, Michail A. Makridis, Nikofors Zacharof, Patrick Langer, Robert Jakob","submitted_at":"2026-05-04T14:34:36Z","abstract_excerpt":"Scientific peer review increasingly struggles to assess reproducibility at the scale and complexity of modern research output. Evaluating reproducibility requires reconstructing experimental dependencies, methodological choices, data flows, and result-generating procedures, which often exceeds what human reviewers can provide. Agentic Reproducibility Assessment (ARA) formalizes reproducibility assessment as a structured reasoning task over scientific documents. Given a paper, ARA extracts a directed workflow graph linking sources, methods, experiments, and outputs, then evaluates its reconstru"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on 213 ReScience C articles demonstrate ARA's generalizability and consistent workflow reconstruction and assessment across LLMs, model temperatures, and scientific domains. ARA achieves ~61% accuracy on three benchmarks, and the highest accuracy reported on ReproBench (60.71% vs. 36.84%) and GoldStandardDB (61.68% vs. 43.56%).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the directed workflow graphs extracted by the agentic system accurately capture the paper's experimental dependencies, data flows, and result-generating procedures in a way that correlates with actual human-validated reproducibility.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"ARA uses LLMs to build workflow graphs linking sources, methods, and outputs in papers, then scores reproducibility, reaching ~61% accuracy on 213 ReScience C articles and outperforming priors on ReproBench and GoldStandardDB.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"ARA extracts directed workflow graphs from papers to evaluate reproducibility at scale.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"657616c77ba0652cac99718b67f6cd776546ed7adf91ad6a1318108d8d87801b"},"source":{"id":"2605.02651","kind":"arxiv","version":2},"verdict":{"id":"adf26f68-68b2-4515-b4f2-49b5c7183e57","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T17:09:47.577891Z","strongest_claim":"Experiments on 213 ReScience C articles demonstrate ARA's generalizability and consistent workflow reconstruction and assessment across LLMs, model temperatures, and scientific domains. ARA achieves ~61% accuracy on three benchmarks, and the highest accuracy reported on ReproBench (60.71% vs. 36.84%) and GoldStandardDB (61.68% vs. 43.56%).","one_line_summary":"ARA uses LLMs to build workflow graphs linking sources, methods, and outputs in papers, then scores reproducibility, reaching ~61% accuracy on 213 ReScience C articles and outperforming priors on ReproBench and GoldStandardDB.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the directed workflow graphs extracted by the agentic system accurately capture the paper's experimental dependencies, data flows, and result-generating procedures in a way that correlates with actual human-validated reproducibility.","pith_extraction_headline":"ARA extracts directed workflow graphs from papers to evaluate reproducibility at scale."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.02651/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_compliance","ran_at":"2026-05-19T16:08:58.816968Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"228d0ce0dc09f8d7a3660bfbaf74045a1a554c2c33c4cafe722f0501b913590a"},"references":{"count":62,"sample":[{"doi":"10.2307/1290335","year":2000,"title":"Publish or perish","work_id":"645bdc08-04c9-4c7c-978f-b9503dffe8e1","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1038/440413a","year":2006,"title":"Science in an exponential world","work_id":"6d117d26-eaeb-4fa7-bb39-cdd415b36894","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.18653/v1/2023.trustnlp-1.17","year":2023,"title":"Distinguishing fact from fiction: A benchmark dataset for identifying machine-generated scientific papers in the llm era","work_id":"3d37b1b4-088e-49ae-a71c-4b42f8f881fb","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1101/2024.03.25.586710","year":2024,"title":"Have ai-generated texts from llm infiltrated the realm of sci- entific writing? a large-scale analysis of preprint platforms","work_id":"6c2fddeb-b9f4-4665-875e-34f6bff5eae1","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Is LLM a reliable reviewer? a comprehensive evaluation of LLM on automatic paper reviewing tasks,","work_id":"30ca2493-b3c7-4a16-8bf5-08d8ad266893","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":62,"snapshot_sha256":"ee36ecdb7b1d422225e4aa365477b64e7034046724de4128e645bf3dff4434c9","internal_anchors":2},"formal_canon":{"evidence_count":2,"snapshot_sha256":"6f0255288638567005bb0bf2ff23e9bf4033195601cd38394da5a273cdcd2bae"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.02651","created_at":"2026-05-20T00:00:40.405525+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.02651v2","created_at":"2026-05-20T00:00:40.405525+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.02651","created_at":"2026-05-20T00:00:40.405525+00:00"},{"alias_kind":"pith_short_12","alias_value":"SNWUXK5XLHWK","created_at":"2026-05-20T00:00:40.405525+00:00"},{"alias_kind":"pith_short_16","alias_value":"SNWUXK5XLHWKECHT","created_at":"2026-05-20T00:00:40.405525+00:00"},{"alias_kind":"pith_short_8","alias_value":"SNWUXK5X","created_at":"2026-05-20T00:00:40.405525+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7","json":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7.json","graph_json":"https://pith.science/api/pith-number/SNWUXK5XLHWKECHTP6XQ7RCCE7/graph.json","events_json":"https://pith.science/api/pith-number/SNWUXK5XLHWKECHTP6XQ7RCCE7/events.json","paper":"https://pith.science/paper/SNWUXK5X"},"agent_actions":{"view_html":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7","download_json":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7.json","view_paper":"https://pith.science/paper/SNWUXK5X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.02651&json=true","fetch_graph":"https://pith.science/api/pith-number/SNWUXK5XLHWKECHTP6XQ7RCCE7/graph.json","fetch_events":"https://pith.science/api/pith-number/SNWUXK5XLHWKECHTP6XQ7RCCE7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7/action/storage_attestation","attest_author":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7/action/author_attestation","sign_citation":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7/action/citation_signature","submit_replication":"https://pith.science/pith/SNWUXK5XLHWKECHTP6XQ7RCCE7/action/replication_record"}},"created_at":"2026-05-20T00:00:40.405525+00:00","updated_at":"2026-05-20T00:00:40.405525+00:00"}