{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:PKSE4PPAV4CZKFGKINVZ66ODTF","short_pith_number":"pith:PKSE4PPA","schema_version":"1.0","canonical_sha256":"7aa44e3de0af059514ca436b9f79c39944601e16c5cb5be39559337f63c0ee75","source":{"kind":"arxiv","id":"1911.03898","version":1},"attestation_state":"computed","paper":{"title":"Understanding Multi-Head Attention in Abstractive Summarization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Anne Schuth, Joris Baan, Maarten de Rijke, Maartje ter Hoeve, Marlies van der Wees","submitted_at":"2019-11-10T10:56:10Z","abstract_excerpt":"Attention mechanisms in deep learning architectures have often been used as a means of transparency and, as such, to shed light on the inner workings of the architectures. Recently, there has been a growing interest in whether or not this assumption is correct. In this paper we investigate the interpretability of multi-head attention in abstractive summarization, a sequence-to-sequence task for which attention does not have an intuitive alignment role, such as in machine translation. We first introduce three metrics to gain insight in the focus of attention heads and observe that these heads s"},"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":false,"formal_links_present":false},"canonical_record":{"source":{"id":"1911.03898","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-11-10T10:56:10Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"1794356cbfbf63a818700144f1cdac93fd7954204850b173b8cfb83bce4638ff","abstract_canon_sha256":"e25bf16d56d68a994d560710df03f83cd12d1392af55c94eded92997e0bb2042"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:18:16.771404Z","signature_b64":"OQOtMlMeVUCjH8qNcVEAT7dl8Kn0CzhEvkB1yihVH6yDRqKy1HAhovQUw5HfXnIOD0j80tZgSdvdHocDzfADDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7aa44e3de0af059514ca436b9f79c39944601e16c5cb5be39559337f63c0ee75","last_reissued_at":"2026-07-05T00:18:16.771003Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:18:16.771003Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding Multi-Head Attention in Abstractive Summarization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Anne Schuth, Joris Baan, Maarten de Rijke, Maartje ter Hoeve, Marlies van der Wees","submitted_at":"2019-11-10T10:56:10Z","abstract_excerpt":"Attention mechanisms in deep learning architectures have often been used as a means of transparency and, as such, to shed light on the inner workings of the architectures. Recently, there has been a growing interest in whether or not this assumption is correct. In this paper we investigate the interpretability of multi-head attention in abstractive summarization, a sequence-to-sequence task for which attention does not have an intuitive alignment role, such as in machine translation. We first introduce three metrics to gain insight in the focus of attention heads and observe that these heads s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1911.03898","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1911.03898/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","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":"1911.03898","created_at":"2026-07-05T00:18:16.771066+00:00"},{"alias_kind":"arxiv_version","alias_value":"1911.03898v1","created_at":"2026-07-05T00:18:16.771066+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1911.03898","created_at":"2026-07-05T00:18:16.771066+00:00"},{"alias_kind":"pith_short_12","alias_value":"PKSE4PPAV4CZ","created_at":"2026-07-05T00:18:16.771066+00:00"},{"alias_kind":"pith_short_16","alias_value":"PKSE4PPAV4CZKFGK","created_at":"2026-07-05T00:18:16.771066+00:00"},{"alias_kind":"pith_short_8","alias_value":"PKSE4PPA","created_at":"2026-07-05T00:18:16.771066+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/PKSE4PPAV4CZKFGKINVZ66ODTF","json":"https://pith.science/pith/PKSE4PPAV4CZKFGKINVZ66ODTF.json","graph_json":"https://pith.science/api/pith-number/PKSE4PPAV4CZKFGKINVZ66ODTF/graph.json","events_json":"https://pith.science/api/pith-number/PKSE4PPAV4CZKFGKINVZ66ODTF/events.json","paper":"https://pith.science/paper/PKSE4PPA"},"agent_actions":{"view_html":"https://pith.science/pith/PKSE4PPAV4CZKFGKINVZ66ODTF","download_json":"https://pith.science/pith/PKSE4PPAV4CZKFGKINVZ66ODTF.json","view_paper":"https://pith.science/paper/PKSE4PPA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1911.03898&json=true","fetch_graph":"https://pith.science/api/pith-number/PKSE4PPAV4CZKFGKINVZ66ODTF/graph.json","fetch_events":"https://pith.science/api/pith-number/PKSE4PPAV4CZKFGKINVZ66ODTF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PKSE4PPAV4CZKFGKINVZ66ODTF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PKSE4PPAV4CZKFGKINVZ66ODTF/action/storage_attestation","attest_author":"https://pith.science/pith/PKSE4PPAV4CZKFGKINVZ66ODTF/action/author_attestation","sign_citation":"https://pith.science/pith/PKSE4PPAV4CZKFGKINVZ66ODTF/action/citation_signature","submit_replication":"https://pith.science/pith/PKSE4PPAV4CZKFGKINVZ66ODTF/action/replication_record"}},"created_at":"2026-07-05T00:18:16.771066+00:00","updated_at":"2026-07-05T00:18:16.771066+00:00"}