{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:EAYAS546LTSUDJZJXO3S5ZXORR","short_pith_number":"pith:EAYAS546","schema_version":"1.0","canonical_sha256":"203009779e5ce541a729bbb72ee6ee8c6f188875b59f54fc6f6ec340fbc5a64e","source":{"kind":"arxiv","id":"2605.09421","version":3},"attestation_state":"computed","paper":{"title":"MACAA: Belief-Revision Multi-Agent Reasoning for Code Authorship Verification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A coordinator and four expert agents use belief revision to verify code authorship without any training data.","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Chenbin Su, Cong Gao, Ge Chu, Jianfei Tang, Jieshuai Yang, Jingwei Ye, Xin Li, Zhi Wang","submitted_at":"2026-05-10T08:47:05Z","abstract_excerpt":"Code authorship attribution (CAA) supports software forensics, plagiarism detection, and intellectual property protection. However, existing supervised CAA approaches suffer from scarce training data and closed-world assumptions: they require sufficient labeled code from fixed candidate-author sets, making training difficult in low-data cases and predictions unreliable for open-world test pairs with unseen samples, or heterogeneous code pairs. Large language models remove task-specific training, but direct prompting depends on costly expert-designed prompts, can hallucinate over complex hetero"},"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":false},"canonical_record":{"source":{"id":"2605.09421","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-05-10T08:47:05Z","cross_cats_sorted":[],"title_canon_sha256":"f7d1c5e72f1d5028cadc87adbfd56e9bb36a97103aea04a9ab30b730904089b6","abstract_canon_sha256":"d883bc926ef9f42dda616c01f1da4d0ac0fa8fb50b7b2b6513e02c4ef9457477"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:43.427096Z","signature_b64":"8BinWHNeVSfgZ3wpFmm1iR1EgMvCD7XOiViupOgfJvrLtbdMTTwRCLc78jFJvLBB8BRfKi1hhUkf9w7Eue7yDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"203009779e5ce541a729bbb72ee6ee8c6f188875b59f54fc6f6ec340fbc5a64e","last_reissued_at":"2026-05-20T00:01:43.426373Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:43.426373Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MACAA: Belief-Revision Multi-Agent Reasoning for Code Authorship Verification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A coordinator and four expert agents use belief revision to verify code authorship without any training data.","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Chenbin Su, Cong Gao, Ge Chu, Jianfei Tang, Jieshuai Yang, Jingwei Ye, Xin Li, Zhi Wang","submitted_at":"2026-05-10T08:47:05Z","abstract_excerpt":"Code authorship attribution (CAA) supports software forensics, plagiarism detection, and intellectual property protection. However, existing supervised CAA approaches suffer from scarce training data and closed-world assumptions: they require sufficient labeled code from fixed candidate-author sets, making training difficult in low-data cases and predictions unreliable for open-world test pairs with unseen samples, or heterogeneous code pairs. Large language models remove task-specific training, but direct prompting depends on costly expert-designed prompts, can hallucinate over complex hetero"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MACAA achieves 89.15% F1 on same-language benchmarks and 80.00% on mixed cross-language pairs, outperforming the baselines overall in both same-language and cross-language evaluations.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The four expert agents reliably extract non-hallucinated evidence from layout, lexical, syntactic, and programming-pattern dimensions even on heterogeneous or cross-language code pairs, and the coordinator's expansion-contraction-revision process produces auditable yet accurate authorship decisions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MACAA is a belief-revision multi-agent framework for training-free code authorship verification that reports 89.15% F1 on same-language benchmarks and 80% on cross-language pairs while outperforming baselines.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A coordinator and four expert agents use belief revision to verify code authorship without any training data.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"b8d9135343bd6443e5b10afaee822bad9030b37e84a97b4d9881e910741ef8b8"},"source":{"id":"2605.09421","kind":"arxiv","version":3},"verdict":{"id":"d7cb5871-1fe4-4cfc-a595-b5794f75a24b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T16:48:57.648527Z","strongest_claim":"MACAA achieves 89.15% F1 on same-language benchmarks and 80.00% on mixed cross-language pairs, outperforming the baselines overall in both same-language and cross-language evaluations.","one_line_summary":"MACAA is a belief-revision multi-agent framework for training-free code authorship verification that reports 89.15% F1 on same-language benchmarks and 80% on cross-language pairs while outperforming baselines.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The four expert agents reliably extract non-hallucinated evidence from layout, lexical, syntactic, and programming-pattern dimensions even on heterogeneous or cross-language code pairs, and the coordinator's expansion-contraction-revision process produces auditable yet accurate authorship decisions.","pith_extraction_headline":"A coordinator and four expert agents use belief revision to verify code authorship without any training data."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.09421/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T19:35:49.560290Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T13:01:18.230409Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T10:12:51.634220Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"ec1389e3bbb256490d11065696eca9213e25be4870c718d730a65ef714e9405b"},"references":{"count":57,"sample":[{"doi":"","year":2021,"title":"Ruchir Puri, David Kung, Geert Janssen, Wei Zhang, Gi- acomo Domeniconi, Vladimir Zolotov, Julian Dolby, Jie Chen, Mihir Choudhury, Lindsey Decker, and 1 others","work_id":"d1085535-e420-42f6-9ee2-8ead4e063167","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"arXiv:2509.05396 [cs]","work_id":"56e70929-251f-4dfe-b21b-83787255f30d","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"naming: both codes use short, lowercase-dominant identifiers (avg_len<3), suggesting consistent personal naming compression habit","work_id":"4619a3fd-3c5f-4945-9bc8-3b4eb1c6c82c","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"structure: both adopt flat, single-block scripts without helper functions or abstractions","work_id":"967c63af-67c3-433c-a615-4de00f492f5f","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"comment: both are comment-free, aligning with rapid-competition authoring style","work_id":"e79f69a5-7312-4bbb-a345-26dd15702563","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":57,"snapshot_sha256":"50806547457b6b5ec25c6baf9b0514c62947463c093f099223606a1107ce41ec","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"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":"2605.09421","created_at":"2026-05-20T00:01:43.426484+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.09421v3","created_at":"2026-05-20T00:01:43.426484+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.09421","created_at":"2026-05-20T00:01:43.426484+00:00"},{"alias_kind":"pith_short_12","alias_value":"EAYAS546LTSU","created_at":"2026-05-20T00:01:43.426484+00:00"},{"alias_kind":"pith_short_16","alias_value":"EAYAS546LTSUDJZJ","created_at":"2026-05-20T00:01:43.426484+00:00"},{"alias_kind":"pith_short_8","alias_value":"EAYAS546","created_at":"2026-05-20T00:01:43.426484+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR","json":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR.json","graph_json":"https://pith.science/api/pith-number/EAYAS546LTSUDJZJXO3S5ZXORR/graph.json","events_json":"https://pith.science/api/pith-number/EAYAS546LTSUDJZJXO3S5ZXORR/events.json","paper":"https://pith.science/paper/EAYAS546"},"agent_actions":{"view_html":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR","download_json":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR.json","view_paper":"https://pith.science/paper/EAYAS546","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.09421&json=true","fetch_graph":"https://pith.science/api/pith-number/EAYAS546LTSUDJZJXO3S5ZXORR/graph.json","fetch_events":"https://pith.science/api/pith-number/EAYAS546LTSUDJZJXO3S5ZXORR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR/action/storage_attestation","attest_author":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR/action/author_attestation","sign_citation":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR/action/citation_signature","submit_replication":"https://pith.science/pith/EAYAS546LTSUDJZJXO3S5ZXORR/action/replication_record"}},"created_at":"2026-05-20T00:01:43.426484+00:00","updated_at":"2026-05-20T00:01:43.426484+00:00"}