{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:NJVM4EMZVBI7YDV6KWFEVQK5X6","short_pith_number":"pith:NJVM4EMZ","schema_version":"1.0","canonical_sha256":"6a6ace1199a851fc0ebe558a4ac15dbf939456a145e092a7c7edeac05937b3f9","source":{"kind":"arxiv","id":"2102.09206","version":3},"attestation_state":"computed","paper":{"title":"Less is More: Pre-train a Strong Text Encoder for Dense Retrieval Using a Weak Decoder","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Arnold Overwijk, Chenyan Xiong, Di He, Guolin Ke, Paul Bennett, Shuqi Lu, Tieyan Liu, Waleed Malik, Zhicheng Dou","submitted_at":"2021-02-18T08:08:17Z","abstract_excerpt":"Dense retrieval requires high-quality text sequence embeddings to support effective search in the representation space. Autoencoder-based language models are appealing in dense retrieval as they train the encoder to output high-quality embedding that can reconstruct the input texts. However, in this paper, we provide theoretical analyses and show empirically that an autoencoder language model with a low reconstruction loss may not provide good sequence representations because the decoder may take shortcuts by exploiting language patterns. To address this, we propose a new self-learning method "},"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":"2102.09206","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-02-18T08:08:17Z","cross_cats_sorted":[],"title_canon_sha256":"57d0e2fc16db409ce1b1f71613447b3f331e2ef40aa64aeb1a6ac8f1550a024e","abstract_canon_sha256":"65cd1ff6c7c31a4cbdcd0a546b52ca8aed19e6ee44088ae40657471ce1a86656"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:14:53.858663Z","signature_b64":"0o4FBnErQ1IphbaKnusPrm6s/8LGDNdnsO9tqmmFLpW4C0c1kMNVobu+NkKtPjo3gDFQ3bTK2PmeHWn0n6K+Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6a6ace1199a851fc0ebe558a4ac15dbf939456a145e092a7c7edeac05937b3f9","last_reissued_at":"2026-07-05T03:14:53.858124Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:14:53.858124Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Less is More: Pre-train a Strong Text Encoder for Dense Retrieval Using a Weak Decoder","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Arnold Overwijk, Chenyan Xiong, Di He, Guolin Ke, Paul Bennett, Shuqi Lu, Tieyan Liu, Waleed Malik, Zhicheng Dou","submitted_at":"2021-02-18T08:08:17Z","abstract_excerpt":"Dense retrieval requires high-quality text sequence embeddings to support effective search in the representation space. Autoencoder-based language models are appealing in dense retrieval as they train the encoder to output high-quality embedding that can reconstruct the input texts. However, in this paper, we provide theoretical analyses and show empirically that an autoencoder language model with a low reconstruction loss may not provide good sequence representations because the decoder may take shortcuts by exploiting language patterns. To address this, we propose a new self-learning method "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.09206","kind":"arxiv","version":3},"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/2102.09206/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":"2102.09206","created_at":"2026-07-05T03:14:53.858189+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.09206v3","created_at":"2026-07-05T03:14:53.858189+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.09206","created_at":"2026-07-05T03:14:53.858189+00:00"},{"alias_kind":"pith_short_12","alias_value":"NJVM4EMZVBI7","created_at":"2026-07-05T03:14:53.858189+00:00"},{"alias_kind":"pith_short_16","alias_value":"NJVM4EMZVBI7YDV6","created_at":"2026-07-05T03:14:53.858189+00:00"},{"alias_kind":"pith_short_8","alias_value":"NJVM4EMZ","created_at":"2026-07-05T03:14:53.858189+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6","json":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6.json","graph_json":"https://pith.science/api/pith-number/NJVM4EMZVBI7YDV6KWFEVQK5X6/graph.json","events_json":"https://pith.science/api/pith-number/NJVM4EMZVBI7YDV6KWFEVQK5X6/events.json","paper":"https://pith.science/paper/NJVM4EMZ"},"agent_actions":{"view_html":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6","download_json":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6.json","view_paper":"https://pith.science/paper/NJVM4EMZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.09206&json=true","fetch_graph":"https://pith.science/api/pith-number/NJVM4EMZVBI7YDV6KWFEVQK5X6/graph.json","fetch_events":"https://pith.science/api/pith-number/NJVM4EMZVBI7YDV6KWFEVQK5X6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6/action/storage_attestation","attest_author":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6/action/author_attestation","sign_citation":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6/action/citation_signature","submit_replication":"https://pith.science/pith/NJVM4EMZVBI7YDV6KWFEVQK5X6/action/replication_record"}},"created_at":"2026-07-05T03:14:53.858189+00:00","updated_at":"2026-07-05T03:14:53.858189+00:00"}