{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:PTTHWPORDM7MOQLQLZYFJ7SSGI","short_pith_number":"pith:PTTHWPOR","schema_version":"1.0","canonical_sha256":"7ce67b3dd11b3ec741705e7054fe5232007e341dfa192ac52afb7a0799bc7c2f","source":{"kind":"arxiv","id":"2010.11386","version":1},"attestation_state":"computed","paper":{"title":"Distilling Dense Representations for Ranking using Tightly-Coupled Teachers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.IR","authors_text":"Jheng-Hong Yang, Jimmy Lin, Sheng-Chieh Lin","submitted_at":"2020-10-22T02:26:01Z","abstract_excerpt":"We present an approach to ranking with dense representations that applies knowledge distillation to improve the recently proposed late-interaction ColBERT model. Specifically, we distill the knowledge from ColBERT's expressive MaxSim operator for computing relevance scores into a simple dot product, thus enabling single-step ANN search. Our key insight is that during distillation, tight coupling between the teacher model and the student model enables more flexible distillation strategies and yields better learned representations. We empirically show that our approach improves query latency and"},"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":"2010.11386","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2020-10-22T02:26:01Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"1cd78233236f52d0a494a7fed8a53b0827f0532423a6b3358696e20f6bd46fc9","abstract_canon_sha256":"55362e35b399edb8ea5b98c883888a0f2f80c73deca494c0615d7998d13917ed"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:45:09.575787Z","signature_b64":"9a+t1qxW6Tns+/vLW52iefH5h5ahaUSj9VpDlOqjwGZbrQW79xrGWlQ5fZJhJxW6N111/b+E5POclCnShlmiBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7ce67b3dd11b3ec741705e7054fe5232007e341dfa192ac52afb7a0799bc7c2f","last_reissued_at":"2026-07-05T01:45:09.575275Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:45:09.575275Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Distilling Dense Representations for Ranking using Tightly-Coupled Teachers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.IR","authors_text":"Jheng-Hong Yang, Jimmy Lin, Sheng-Chieh Lin","submitted_at":"2020-10-22T02:26:01Z","abstract_excerpt":"We present an approach to ranking with dense representations that applies knowledge distillation to improve the recently proposed late-interaction ColBERT model. Specifically, we distill the knowledge from ColBERT's expressive MaxSim operator for computing relevance scores into a simple dot product, thus enabling single-step ANN search. Our key insight is that during distillation, tight coupling between the teacher model and the student model enables more flexible distillation strategies and yields better learned representations. We empirically show that our approach improves query latency and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.11386","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/2010.11386/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":"2010.11386","created_at":"2026-07-05T01:45:09.575343+00:00"},{"alias_kind":"arxiv_version","alias_value":"2010.11386v1","created_at":"2026-07-05T01:45:09.575343+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.11386","created_at":"2026-07-05T01:45:09.575343+00:00"},{"alias_kind":"pith_short_12","alias_value":"PTTHWPORDM7M","created_at":"2026-07-05T01:45:09.575343+00:00"},{"alias_kind":"pith_short_16","alias_value":"PTTHWPORDM7MOQLQ","created_at":"2026-07-05T01:45:09.575343+00:00"},{"alias_kind":"pith_short_8","alias_value":"PTTHWPOR","created_at":"2026-07-05T01:45:09.575343+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03244","citing_title":"When Does Complexity Conditioning Help a Frozen Sentence Embedding? A Controlled Study of Per-Sentence and Pair-Level Difficulty Adaptation","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07930","citing_title":"Unified Supervision for Walmart's Sponsored Search Retrieval via Joint Semantic Relevance and Behavioral Engagement Modeling","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI","json":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI.json","graph_json":"https://pith.science/api/pith-number/PTTHWPORDM7MOQLQLZYFJ7SSGI/graph.json","events_json":"https://pith.science/api/pith-number/PTTHWPORDM7MOQLQLZYFJ7SSGI/events.json","paper":"https://pith.science/paper/PTTHWPOR"},"agent_actions":{"view_html":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI","download_json":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI.json","view_paper":"https://pith.science/paper/PTTHWPOR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2010.11386&json=true","fetch_graph":"https://pith.science/api/pith-number/PTTHWPORDM7MOQLQLZYFJ7SSGI/graph.json","fetch_events":"https://pith.science/api/pith-number/PTTHWPORDM7MOQLQLZYFJ7SSGI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI/action/storage_attestation","attest_author":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI/action/author_attestation","sign_citation":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI/action/citation_signature","submit_replication":"https://pith.science/pith/PTTHWPORDM7MOQLQLZYFJ7SSGI/action/replication_record"}},"created_at":"2026-07-05T01:45:09.575343+00:00","updated_at":"2026-07-05T01:45:09.575343+00:00"}