{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:V7O7UVWBFKOB4RXSX2OQHYWRDJ","short_pith_number":"pith:V7O7UVWB","schema_version":"1.0","canonical_sha256":"afddfa56c12a9c1e46f2be9d03e2d11a66a57b99f09e8838a73b43fb18a1c7cb","source":{"kind":"arxiv","id":"2210.11773","version":2},"attestation_state":"computed","paper":{"title":"SimANS: Simple Ambiguous Negatives Sampling for Dense Text Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IR"],"primary_cat":"cs.CL","authors_text":"Anlei Dong, Jingwen Lu, Ji-Rong Wen, Kun Zhou, Nan Duan, Rangan Majumder, Wayne Xin Zhao, Weizhu Chen, Xiao Liu, Yelong Shen, Yeyun Gong","submitted_at":"2022-10-21T07:18:05Z","abstract_excerpt":"Sampling proper negatives from a large document pool is vital to effectively train a dense retrieval model. However, existing negative sampling strategies suffer from the uninformative or false negative problem. In this work, we empirically show that according to the measured relevance scores, the negatives ranked around the positives are generally more informative and less likely to be false negatives. Intuitively, these negatives are not too hard (\\emph{may be false negatives}) or too easy (\\emph{uninformative}). They are the ambiguous negatives and need more attention during training. Thus,"},"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":"2210.11773","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-10-21T07:18:05Z","cross_cats_sorted":["cs.IR"],"title_canon_sha256":"057f1feb4ea3329cce13aae1848f010da9cc11a200fbec3e8fb8e28b1d13a07d","abstract_canon_sha256":"05736062cc3f47e3ca3c1ff9aaf78ba2baeab82a825685fb70131773e08e4367"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:09:24.126150Z","signature_b64":"HU0hEqrfltC0rsF6DmvRD5lx7wnGjw5Eq8F501rCRwutyhky1aXWzby6aumz1lzfnQZKm+QqEST0tJy5XYYdDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"afddfa56c12a9c1e46f2be9d03e2d11a66a57b99f09e8838a73b43fb18a1c7cb","last_reissued_at":"2026-07-05T05:09:24.125614Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:09:24.125614Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SimANS: Simple Ambiguous Negatives Sampling for Dense Text Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.IR"],"primary_cat":"cs.CL","authors_text":"Anlei Dong, Jingwen Lu, Ji-Rong Wen, Kun Zhou, Nan Duan, Rangan Majumder, Wayne Xin Zhao, Weizhu Chen, Xiao Liu, Yelong Shen, Yeyun Gong","submitted_at":"2022-10-21T07:18:05Z","abstract_excerpt":"Sampling proper negatives from a large document pool is vital to effectively train a dense retrieval model. However, existing negative sampling strategies suffer from the uninformative or false negative problem. In this work, we empirically show that according to the measured relevance scores, the negatives ranked around the positives are generally more informative and less likely to be false negatives. Intuitively, these negatives are not too hard (\\emph{may be false negatives}) or too easy (\\emph{uninformative}). They are the ambiguous negatives and need more attention during training. Thus,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.11773","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/2210.11773/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":"2210.11773","created_at":"2026-07-05T05:09:24.125682+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.11773v2","created_at":"2026-07-05T05:09:24.125682+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.11773","created_at":"2026-07-05T05:09:24.125682+00:00"},{"alias_kind":"pith_short_12","alias_value":"V7O7UVWBFKOB","created_at":"2026-07-05T05:09:24.125682+00:00"},{"alias_kind":"pith_short_16","alias_value":"V7O7UVWBFKOB4RXS","created_at":"2026-07-05T05:09:24.125682+00:00"},{"alias_kind":"pith_short_8","alias_value":"V7O7UVWB","created_at":"2026-07-05T05:09:24.125682+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2401.03563","citing_title":"Data-CUBE: Data Curriculum for Instruction-based Sentence Representation Learning","ref_index":55,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ","json":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ.json","graph_json":"https://pith.science/api/pith-number/V7O7UVWBFKOB4RXSX2OQHYWRDJ/graph.json","events_json":"https://pith.science/api/pith-number/V7O7UVWBFKOB4RXSX2OQHYWRDJ/events.json","paper":"https://pith.science/paper/V7O7UVWB"},"agent_actions":{"view_html":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ","download_json":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ.json","view_paper":"https://pith.science/paper/V7O7UVWB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.11773&json=true","fetch_graph":"https://pith.science/api/pith-number/V7O7UVWBFKOB4RXSX2OQHYWRDJ/graph.json","fetch_events":"https://pith.science/api/pith-number/V7O7UVWBFKOB4RXSX2OQHYWRDJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ/action/storage_attestation","attest_author":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ/action/author_attestation","sign_citation":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ/action/citation_signature","submit_replication":"https://pith.science/pith/V7O7UVWBFKOB4RXSX2OQHYWRDJ/action/replication_record"}},"created_at":"2026-07-05T05:09:24.125682+00:00","updated_at":"2026-07-05T05:09:24.125682+00:00"}