{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:7EZAARD7W4TRTFCPFJU357PEDJ","short_pith_number":"pith:7EZAARD7","schema_version":"1.0","canonical_sha256":"f93200447fb72719944f2a69befde41a6758832124f5eaf52b5ec557d91f9022","source":{"kind":"arxiv","id":"2302.07452","version":1},"attestation_state":"computed","paper":{"title":"How to Train Your DRAGON: Diverse Augmentation Towards Generalizable Dense Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.IR","authors_text":"Akari Asai, Barlas Oguz, Jimmy Lin, Minghan Li, Sheng-Chieh Lin, Wen-tau Yih, Xilun Chen, Yashar Mehdad","submitted_at":"2023-02-15T03:53:26Z","abstract_excerpt":"Various techniques have been developed in recent years to improve dense retrieval (DR), such as unsupervised contrastive learning and pseudo-query generation. Existing DRs, however, often suffer from effectiveness tradeoffs between supervised and zero-shot retrieval, which some argue was due to the limited model capacity. We contradict this hypothesis and show that a generalizable DR can be trained to achieve high accuracy in both supervised and zero-shot retrieval without increasing model size. In particular, we systematically examine the contrastive learning of DRs, under the framework of Da"},"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":"2302.07452","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.IR","submitted_at":"2023-02-15T03:53:26Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"364ba302d9ef64326496777650291edd490334ec83930bc9f527897529a8224e","abstract_canon_sha256":"bca4ac516a2df47b9c74d0cea27c17bb9ad667009a8b2a26d93cfcb415aafeaa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:41:54.220466Z","signature_b64":"fXDnUNlwf/f+FQWlWnDGUpI/pxnX7OlJuua9gG3ES3v1HHu9AEf21howfcVEwtlhYhUQr8S6K3TnuPu1bdBVDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f93200447fb72719944f2a69befde41a6758832124f5eaf52b5ec557d91f9022","last_reissued_at":"2026-07-05T05:41:54.220034Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:41:54.220034Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How to Train Your DRAGON: Diverse Augmentation Towards Generalizable Dense Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.IR","authors_text":"Akari Asai, Barlas Oguz, Jimmy Lin, Minghan Li, Sheng-Chieh Lin, Wen-tau Yih, Xilun Chen, Yashar Mehdad","submitted_at":"2023-02-15T03:53:26Z","abstract_excerpt":"Various techniques have been developed in recent years to improve dense retrieval (DR), such as unsupervised contrastive learning and pseudo-query generation. Existing DRs, however, often suffer from effectiveness tradeoffs between supervised and zero-shot retrieval, which some argue was due to the limited model capacity. We contradict this hypothesis and show that a generalizable DR can be trained to achieve high accuracy in both supervised and zero-shot retrieval without increasing model size. In particular, we systematically examine the contrastive learning of DRs, under the framework of Da"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.07452","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/2302.07452/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":"2302.07452","created_at":"2026-07-05T05:41:54.220093+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.07452v1","created_at":"2026-07-05T05:41:54.220093+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.07452","created_at":"2026-07-05T05:41:54.220093+00:00"},{"alias_kind":"pith_short_12","alias_value":"7EZAARD7W4TR","created_at":"2026-07-05T05:41:54.220093+00:00"},{"alias_kind":"pith_short_16","alias_value":"7EZAARD7W4TRTFCP","created_at":"2026-07-05T05:41:54.220093+00:00"},{"alias_kind":"pith_short_8","alias_value":"7EZAARD7","created_at":"2026-07-05T05:41:54.220093+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2506.04565","citing_title":"From Standalone LLMs to Integrated Intelligence: A Survey of Compound Al Systems","ref_index":94,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07041","citing_title":"AV-SQL: Decomposing Complex Text-to-SQL Queries with Agentic Views","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06163","citing_title":"Data, Not Model: Explaining Bias toward LLM Texts in Neural Retrievers","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04734","citing_title":"Beyond Hard Negatives: The Importance of Score Distribution in Knowledge Distillation for Dense Retrieval","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ","json":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ.json","graph_json":"https://pith.science/api/pith-number/7EZAARD7W4TRTFCPFJU357PEDJ/graph.json","events_json":"https://pith.science/api/pith-number/7EZAARD7W4TRTFCPFJU357PEDJ/events.json","paper":"https://pith.science/paper/7EZAARD7"},"agent_actions":{"view_html":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ","download_json":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ.json","view_paper":"https://pith.science/paper/7EZAARD7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.07452&json=true","fetch_graph":"https://pith.science/api/pith-number/7EZAARD7W4TRTFCPFJU357PEDJ/graph.json","fetch_events":"https://pith.science/api/pith-number/7EZAARD7W4TRTFCPFJU357PEDJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ/action/storage_attestation","attest_author":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ/action/author_attestation","sign_citation":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ/action/citation_signature","submit_replication":"https://pith.science/pith/7EZAARD7W4TRTFCPFJU357PEDJ/action/replication_record"}},"created_at":"2026-07-05T05:41:54.220093+00:00","updated_at":"2026-07-05T05:41:54.220093+00:00"}