{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:4IM2UBJ5KCQCMXTANSJNDR4WYO","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"18ef241549ecb5f59290f6518964006776aa7c47a1c0eaa8a929d1c0b34218be","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-08-29T02:03:05Z","title_canon_sha256":"e80fe4a55a1a4e2af7be4102f45829039a26319d33c38d432dd039d14a676cdd"},"schema_version":"1.0","source":{"id":"2408.16208","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.16208","created_at":"2026-07-05T09:00:27Z"},{"alias_kind":"arxiv_version","alias_value":"2408.16208v1","created_at":"2026-07-05T09:00:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.16208","created_at":"2026-07-05T09:00:27Z"},{"alias_kind":"pith_short_12","alias_value":"4IM2UBJ5KCQC","created_at":"2026-07-05T09:00:27Z"},{"alias_kind":"pith_short_16","alias_value":"4IM2UBJ5KCQCMXTA","created_at":"2026-07-05T09:00:27Z"},{"alias_kind":"pith_short_8","alias_value":"4IM2UBJ5","created_at":"2026-07-05T09:00:27Z"}],"graph_snapshots":[{"event_id":"sha256:3bdeb3589adaedabee39c2d567bd59803a4d8861252b6b83cd5312c974687749","target":"graph","created_at":"2026-07-05T09:00:27Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2408.16208/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Given the rapidly expanding capabilities of generative AI models for radiology, there is a need for robust metrics that can accurately measure the quality of AI-generated radiology reports across diverse hospitals. We develop ReXamine-Global, a LLM-powered, multi-site framework that tests metrics across different writing styles and patient populations, exposing gaps in their generalization. First, our method tests whether a metric is undesirably sensitive to reporting style, providing different scores depending on whether AI-generated reports are stylistically similar to ground-truth reports o","authors_text":"Adil Zia, Agustina Saenz, Alain S. Abi-Ghanem, Cibele Luna, Dominic Buensalido, Frederic Jonske, Helen Kavnoudias, Heng-sheng Chao, Johannes Haubold, Kay Wu, Khaled Al-Surimi, Lars Heiliger, Moon Kim, Nour El Ghawi, Oishi Banerjee, Patricia Castillo, Pranav Rajpurkar, Rayyan A. Daghistani, Warren Clements, Yuh-Min Chen","cross_cats":["cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-08-29T02:03:05Z","title":"ReXamine-Global: A Framework for Uncovering Inconsistencies in Radiology Report Generation Metrics"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.16208","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:bfe1c7027dcf731e6b52d40303aba4cb79be3b08f90dbb79f81f7fdf22b2366a","target":"record","created_at":"2026-07-05T09:00:27Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"18ef241549ecb5f59290f6518964006776aa7c47a1c0eaa8a929d1c0b34218be","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-08-29T02:03:05Z","title_canon_sha256":"e80fe4a55a1a4e2af7be4102f45829039a26319d33c38d432dd039d14a676cdd"},"schema_version":"1.0","source":{"id":"2408.16208","kind":"arxiv","version":1}},"canonical_sha256":"e219aa053d50a0265e606c92d1c796c3a3c7f32be384f5cd2d7b807108acd121","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e219aa053d50a0265e606c92d1c796c3a3c7f32be384f5cd2d7b807108acd121","first_computed_at":"2026-07-05T09:00:27.796759Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:00:27.796759Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"M0pkI30BeDsHwAy+4vR6wO4fVtCWN/SjufzolTmn/k8sReXnhl+MMb3gf6tStzchZRPMri0x1jyizJquC3/XDg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:00:27.797215Z","signed_message":"canonical_sha256_bytes"},"source_id":"2408.16208","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bfe1c7027dcf731e6b52d40303aba4cb79be3b08f90dbb79f81f7fdf22b2366a","sha256:3bdeb3589adaedabee39c2d567bd59803a4d8861252b6b83cd5312c974687749"],"state_sha256":"56f90b77e25d280a2737b59170cbf3ae9e08bd0b444427c2576cee4266d90a53"}