{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:KX2T45ZWBT76TO267MZ557RCRG","short_pith_number":"pith:KX2T45ZW","schema_version":"1.0","canonical_sha256":"55f53e77360cffe9bb5efb33defe2289adbbd6020af15a85291d14e88c68dcf1","source":{"kind":"arxiv","id":"2110.08499","version":2},"attestation_state":"computed","paper":{"title":"PRIMERA: Pyramid-based Masked Sentence Pre-training for Multi-document Summarization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Arman Cohan, Giuseppe Carenini, Iz Beltagy, Wen Xiao","submitted_at":"2021-10-16T07:22:24Z","abstract_excerpt":"We introduce PRIMERA, a pre-trained model for multi-document representation with a focus on summarization that reduces the need for dataset-specific architectures and large amounts of fine-tuning labeled data. PRIMERA uses our newly proposed pre-training objective designed to teach the model to connect and aggregate information across documents. It also uses efficient encoder-decoder transformers to simplify the processing of concatenated input documents. With extensive experiments on 6 multi-document summarization datasets from 3 different domains on zero-shot, few-shot and full-supervised se"},"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":"2110.08499","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-10-16T07:22:24Z","cross_cats_sorted":[],"title_canon_sha256":"cfd75d07bbc121db382994b7ba6610b9b9b720f6a7ea098b3685edbb397138ba","abstract_canon_sha256":"06dce9e438797fdfcc3d1a380e41e6100251e617c80b1e8f2b8555ae0f0fbad9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:06:00.470332Z","signature_b64":"q5Vy0grvCWt0pXNFr1waLx6dD3kU9kt2J9p9FfiGtrp4d/YWOGO7cJgSKGaEV+HUSJXIkVgwd+w8b2QoK34NCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"55f53e77360cffe9bb5efb33defe2289adbbd6020af15a85291d14e88c68dcf1","last_reissued_at":"2026-07-05T04:06:00.469852Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:06:00.469852Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PRIMERA: Pyramid-based Masked Sentence Pre-training for Multi-document Summarization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Arman Cohan, Giuseppe Carenini, Iz Beltagy, Wen Xiao","submitted_at":"2021-10-16T07:22:24Z","abstract_excerpt":"We introduce PRIMERA, a pre-trained model for multi-document representation with a focus on summarization that reduces the need for dataset-specific architectures and large amounts of fine-tuning labeled data. PRIMERA uses our newly proposed pre-training objective designed to teach the model to connect and aggregate information across documents. It also uses efficient encoder-decoder transformers to simplify the processing of concatenated input documents. With extensive experiments on 6 multi-document summarization datasets from 3 different domains on zero-shot, few-shot and full-supervised se"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.08499","kind":"arxiv","version":2},"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/2110.08499/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":"2110.08499","created_at":"2026-07-05T04:06:00.469908+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.08499v2","created_at":"2026-07-05T04:06:00.469908+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.08499","created_at":"2026-07-05T04:06:00.469908+00:00"},{"alias_kind":"pith_short_12","alias_value":"KX2T45ZWBT76","created_at":"2026-07-05T04:06:00.469908+00:00"},{"alias_kind":"pith_short_16","alias_value":"KX2T45ZWBT76TO26","created_at":"2026-07-05T04:06:00.469908+00:00"},{"alias_kind":"pith_short_8","alias_value":"KX2T45ZW","created_at":"2026-07-05T04:06:00.469908+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19591","citing_title":"A BART-based approach with hierarchical strategy for Vietnamese abstractive multi-document summarization","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2404.07143","citing_title":"Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG","json":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG.json","graph_json":"https://pith.science/api/pith-number/KX2T45ZWBT76TO267MZ557RCRG/graph.json","events_json":"https://pith.science/api/pith-number/KX2T45ZWBT76TO267MZ557RCRG/events.json","paper":"https://pith.science/paper/KX2T45ZW"},"agent_actions":{"view_html":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG","download_json":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG.json","view_paper":"https://pith.science/paper/KX2T45ZW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.08499&json=true","fetch_graph":"https://pith.science/api/pith-number/KX2T45ZWBT76TO267MZ557RCRG/graph.json","fetch_events":"https://pith.science/api/pith-number/KX2T45ZWBT76TO267MZ557RCRG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG/action/storage_attestation","attest_author":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG/action/author_attestation","sign_citation":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG/action/citation_signature","submit_replication":"https://pith.science/pith/KX2T45ZWBT76TO267MZ557RCRG/action/replication_record"}},"created_at":"2026-07-05T04:06:00.469908+00:00","updated_at":"2026-07-05T04:06:00.469908+00:00"}