{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UWLAOSEBDCOIRKPW6IPJ2KVYU6","short_pith_number":"pith:UWLAOSEB","schema_version":"1.0","canonical_sha256":"a596074881189c88a9f6f21e9d2ab8a7906a27936f913b88eade7c18287ce21d","source":{"kind":"arxiv","id":"2605.13451","version":1},"attestation_state":"computed","paper":{"title":"LongBEL: Long-Context and Document-Consistent Biomedical Entity Linking","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"LongBEL links biomedical mentions consistently across full documents by maintaining a memory of prior predictions.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Adam Remaki, Christel G\\'erardin, Xavier Tannier","submitted_at":"2026-05-13T12:47:21Z","abstract_excerpt":"Biomedical entity linking maps textual mentions to concepts in structured knowledge bases such as UMLS or SNOMED CT. Most existing systems link each mention independently, using only the mention or its surrounding sentence. This ignores dependencies between mentions in the same document and can lead to inconsistent predictions, especially when the same concept appears under different surface forms. We introduce LongBEL, a document-level generative framework that combines full-document context with a memory of previous predictions. To make this memory robust, LongBEL is trained with cross-valid"},"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":"2605.13451","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-13T12:47:21Z","cross_cats_sorted":[],"title_canon_sha256":"7c31702593aec5c2d5859c88f6b6dfc5ae571af36b60f65d2f6f33dd3da1bce7","abstract_canon_sha256":"31b232ee43c273a5e5b9479b1e24a651900eb7a79a0dba13b0127669b1292a52"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:41.891489Z","signature_b64":"7jnQjsRE/VIfauEvljpCWOvJoYmoSZ876GUdTJlF9vJ1kQOC9m2/8MxUty8n6zgG3VsRz+BSFfBvDzbC5hZ9Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a596074881189c88a9f6f21e9d2ab8a7906a27936f913b88eade7c18287ce21d","last_reissued_at":"2026-05-18T02:44:41.890969Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:41.890969Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LongBEL: Long-Context and Document-Consistent Biomedical Entity Linking","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"LongBEL links biomedical mentions consistently across full documents by maintaining a memory of prior predictions.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Adam Remaki, Christel G\\'erardin, Xavier Tannier","submitted_at":"2026-05-13T12:47:21Z","abstract_excerpt":"Biomedical entity linking maps textual mentions to concepts in structured knowledge bases such as UMLS or SNOMED CT. Most existing systems link each mention independently, using only the mention or its surrounding sentence. This ignores dependencies between mentions in the same document and can lead to inconsistent predictions, especially when the same concept appears under different surface forms. We introduce LongBEL, a document-level generative framework that combines full-document context with a memory of previous predictions. To make this memory robust, LongBEL is trained with cross-valid"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"LongBEL improves over sentence-level generative baselines, with the largest gains on datasets where concepts frequently recur within documents. An ensemble of local, global, and memory-based variants achieves the best results across all benchmarks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That training the memory component exclusively with cross-validated predictions sufficiently reduces train-inference mismatch and prevents cascading errors without introducing new inconsistencies.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LongBEL improves biomedical entity linking consistency by combining full-document context with memory of previous predictions trained via cross-validation rather than gold labels.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"LongBEL links biomedical mentions consistently across full documents by maintaining a memory of prior predictions.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4e8e8b05bbb9aec71d62d70d8ac0dc5a5d5e8964f46aab67c2f6958de5444b8a"},"source":{"id":"2605.13451","kind":"arxiv","version":1},"verdict":{"id":"da635265-5224-4688-a945-05cb1e393c39","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:21:33.168694Z","strongest_claim":"LongBEL improves over sentence-level generative baselines, with the largest gains on datasets where concepts frequently recur within documents. An ensemble of local, global, and memory-based variants achieves the best results across all benchmarks.","one_line_summary":"LongBEL improves biomedical entity linking consistency by combining full-document context with memory of previous predictions trained via cross-validation rather than gold labels.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That training the memory component exclusively with cross-validated predictions sufficiently reduces train-inference mismatch and prevents cascading errors without introducing new inconsistencies.","pith_extraction_headline":"LongBEL links biomedical mentions consistently across full documents by maintaining a memory of prior predictions."},"references":{"count":50,"sample":[{"doi":"10.1016/j.jbi.2022.104252","year":2023,"title":"Journal of biomedical informatics , author =","work_id":"57084228-dc4e-4ac7-bf21-e0d7eb294586","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.2196/68704","year":2025,"title":"JMIR Medical Informatics , author =","work_id":"c0a93395-f86a-463f-98c0-bd1ab3504d58","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.48550/arxiv.2008.00401","year":2008,"title":"Ryan Wong, Necati Cihan Camgoz, and Richard Bow- den","work_id":"8ae4fd91-262a-4425-a73f-59e60c8e1865","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Névéol, Aurélie and Grouin, Cyril and Leixa, Jeremy and Rosset, Sophie and Zweigenbaum, Pierre , year =. The. Proc of","work_id":"ade16ffb-c56c-4a0a-9844-10efd9f75fc5","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Miranda-Escalada, Antonio and Gasco, Luis and Lima-López, Salvador and Farré-Maduell, Eulàlia and Estrada, Darryl and Nentidis, Anastasios and Krithara, Anastasia and Katsimpras, Georgios and Palioura","work_id":"bd349e10-0175-4fc5-8d4a-0c57cef9b916","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":50,"snapshot_sha256":"784ee915478bc86c2ab4ea63e876acaef3e13c25ae06541689ab31591ac19dd8","internal_anchors":3},"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":"2605.13451","created_at":"2026-05-18T02:44:41.891048+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.13451v1","created_at":"2026-05-18T02:44:41.891048+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13451","created_at":"2026-05-18T02:44:41.891048+00:00"},{"alias_kind":"pith_short_12","alias_value":"UWLAOSEBDCOI","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"UWLAOSEBDCOIRKPW","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"UWLAOSEB","created_at":"2026-05-18T12:33:37.589309+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/UWLAOSEBDCOIRKPW6IPJ2KVYU6","json":"https://pith.science/pith/UWLAOSEBDCOIRKPW6IPJ2KVYU6.json","graph_json":"https://pith.science/api/pith-number/UWLAOSEBDCOIRKPW6IPJ2KVYU6/graph.json","events_json":"https://pith.science/api/pith-number/UWLAOSEBDCOIRKPW6IPJ2KVYU6/events.json","paper":"https://pith.science/paper/UWLAOSEB"},"agent_actions":{"view_html":"https://pith.science/pith/UWLAOSEBDCOIRKPW6IPJ2KVYU6","download_json":"https://pith.science/pith/UWLAOSEBDCOIRKPW6IPJ2KVYU6.json","view_paper":"https://pith.science/paper/UWLAOSEB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.13451&json=true","fetch_graph":"https://pith.science/api/pith-number/UWLAOSEBDCOIRKPW6IPJ2KVYU6/graph.json","fetch_events":"https://pith.science/api/pith-number/UWLAOSEBDCOIRKPW6IPJ2KVYU6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UWLAOSEBDCOIRKPW6IPJ2KVYU6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UWLAOSEBDCOIRKPW6IPJ2KVYU6/action/storage_attestation","attest_author":"https://pith.science/pith/UWLAOSEBDCOIRKPW6IPJ2KVYU6/action/author_attestation","sign_citation":"https://pith.science/pith/UWLAOSEBDCOIRKPW6IPJ2KVYU6/action/citation_signature","submit_replication":"https://pith.science/pith/UWLAOSEBDCOIRKPW6IPJ2KVYU6/action/replication_record"}},"created_at":"2026-05-18T02:44:41.891048+00:00","updated_at":"2026-05-18T02:44:41.891048+00:00"}