{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:F7S277WKSXNZRGPD2JVCTZVY6R","short_pith_number":"pith:F7S277WK","schema_version":"1.0","canonical_sha256":"2fe5affeca95db9899e3d26a29e6b8f47c01fc8309c1a02cfcad421f59323fa7","source":{"kind":"arxiv","id":"1905.01969","version":4},"attestation_state":"computed","paper":{"title":"Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jason Weston, Kurt Shuster, Marie-Anne Lachaux, Samuel Humeau","submitted_at":"2019-04-22T02:18:00Z","abstract_excerpt":"The use of deep pre-trained bidirectional transformers has led to remarkable progress in a number of applications (Devlin et al., 2018). For tasks that make pairwise comparisons between sequences, matching a given input with a corresponding label, two approaches are common: Cross-encoders performing full self-attention over the pair and Bi-encoders encoding the pair separately. The former often performs better, but is too slow for practical use. In this work, we develop a new transformer architecture, the Poly-encoder, that learns global rather than token level self-attention features. We perf"},"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":"1905.01969","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-04-22T02:18:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1259f3cc25801c926418de084b8ed6d3a18fcc9e6e706746c0508c93dffe34a7","abstract_canon_sha256":"47f941fd88961f52a185ef4c7dc45c7c4c65f4205db49b8a2f163b8bbbe52ae0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:50:42.980187Z","signature_b64":"IoWwhjMo0+wcYxuSz5QMoD1Oc+SHFHZWgVPtkq7Dda+IvgQZIZ/v341IBXt8YvJJvVGc8orgDeLSzrS58Vt+Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2fe5affeca95db9899e3d26a29e6b8f47c01fc8309c1a02cfcad421f59323fa7","last_reissued_at":"2026-07-05T00:50:42.979725Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:50:42.979725Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jason Weston, Kurt Shuster, Marie-Anne Lachaux, Samuel Humeau","submitted_at":"2019-04-22T02:18:00Z","abstract_excerpt":"The use of deep pre-trained bidirectional transformers has led to remarkable progress in a number of applications (Devlin et al., 2018). For tasks that make pairwise comparisons between sequences, matching a given input with a corresponding label, two approaches are common: Cross-encoders performing full self-attention over the pair and Bi-encoders encoding the pair separately. The former often performs better, but is too slow for practical use. In this work, we develop a new transformer architecture, the Poly-encoder, that learns global rather than token level self-attention features. We perf"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.01969","kind":"arxiv","version":4},"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/1905.01969/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":"1905.01969","created_at":"2026-07-05T00:50:42.979781+00:00"},{"alias_kind":"arxiv_version","alias_value":"1905.01969v4","created_at":"2026-07-05T00:50:42.979781+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.01969","created_at":"2026-07-05T00:50:42.979781+00:00"},{"alias_kind":"pith_short_12","alias_value":"F7S277WKSXNZ","created_at":"2026-07-05T00:50:42.979781+00:00"},{"alias_kind":"pith_short_16","alias_value":"F7S277WKSXNZRGPD","created_at":"2026-07-05T00:50:42.979781+00:00"},{"alias_kind":"pith_short_8","alias_value":"F7S277WK","created_at":"2026-07-05T00:50:42.979781+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05971","citing_title":"Multimodal Video-to-Music Recommendation via Semantic Retrieval and Temporal Reranking","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2605.22379","citing_title":"Cross-Subject EEG Emotion Recognition Based on Temporal Asynchronous Alignment Contrastive Learning","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2208.03299","citing_title":"Atlas: Few-shot Learning with Retrieval Augmented Language Models","ref_index":103,"is_internal_anchor":false},{"citing_arxiv_id":"2112.09118","citing_title":"Unsupervised Dense Information Retrieval with Contrastive Learning","ref_index":138,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05034","citing_title":"Learning to Unscramble Feynman Loop Integrals with SAILIR","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"1908.10084","citing_title":"Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R","json":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R.json","graph_json":"https://pith.science/api/pith-number/F7S277WKSXNZRGPD2JVCTZVY6R/graph.json","events_json":"https://pith.science/api/pith-number/F7S277WKSXNZRGPD2JVCTZVY6R/events.json","paper":"https://pith.science/paper/F7S277WK"},"agent_actions":{"view_html":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R","download_json":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R.json","view_paper":"https://pith.science/paper/F7S277WK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1905.01969&json=true","fetch_graph":"https://pith.science/api/pith-number/F7S277WKSXNZRGPD2JVCTZVY6R/graph.json","fetch_events":"https://pith.science/api/pith-number/F7S277WKSXNZRGPD2JVCTZVY6R/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R/action/storage_attestation","attest_author":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R/action/author_attestation","sign_citation":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R/action/citation_signature","submit_replication":"https://pith.science/pith/F7S277WKSXNZRGPD2JVCTZVY6R/action/replication_record"}},"created_at":"2026-07-05T00:50:42.979781+00:00","updated_at":"2026-07-05T00:50:42.979781+00:00"}