{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:R2YIHYELNB7CYBYO3CZABQLLAK","short_pith_number":"pith:R2YIHYEL","schema_version":"1.0","canonical_sha256":"8eb083e08b687e2c070ed8b200c16b02ae1885744517f8db9c749a14d600e7f8","source":{"kind":"arxiv","id":"2104.05883","version":1},"attestation_state":"computed","paper":{"title":"Multi-Step Reasoning Over Unstructured Text with Beam Dense Retrieval","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chenyan Xiong, Chen Zhao, Hal Daum\\'e III, Jordan Boyd-Graber","submitted_at":"2021-04-13T01:16:48Z","abstract_excerpt":"Complex question answering often requires finding a reasoning chain that consists of multiple evidence pieces. Current approaches incorporate the strengths of structured knowledge and unstructured text, assuming text corpora is semi-structured. Building on dense retrieval methods, we propose a new multi-step retrieval approach (BeamDR) that iteratively forms an evidence chain through beam search in dense representations. When evaluated on multi-hop question answering, BeamDR is competitive to state-of-the-art systems, without using any semi-structured information. Through query composition in "},"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":"2104.05883","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-04-13T01:16:48Z","cross_cats_sorted":[],"title_canon_sha256":"d780aca5165f4729a542103875c5f04576d92916aedf16a4b0d1abf2e6e5dc97","abstract_canon_sha256":"30a9f83f8ee940d3c9d78a47c2990a96da2f69e88ee39e032ca84eff04833493"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:31:26.258094Z","signature_b64":"O9NUQCGYrKFTutpQUbbR6wonR97USYX7x2LHraoLQKP9w3PcavXkL8q3mmutwcKtfBWFAQrIuvi7sDO0sDm0CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8eb083e08b687e2c070ed8b200c16b02ae1885744517f8db9c749a14d600e7f8","last_reissued_at":"2026-07-05T02:31:26.257632Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:31:26.257632Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multi-Step Reasoning Over Unstructured Text with Beam Dense Retrieval","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chenyan Xiong, Chen Zhao, Hal Daum\\'e III, Jordan Boyd-Graber","submitted_at":"2021-04-13T01:16:48Z","abstract_excerpt":"Complex question answering often requires finding a reasoning chain that consists of multiple evidence pieces. Current approaches incorporate the strengths of structured knowledge and unstructured text, assuming text corpora is semi-structured. Building on dense retrieval methods, we propose a new multi-step retrieval approach (BeamDR) that iteratively forms an evidence chain through beam search in dense representations. When evaluated on multi-hop question answering, BeamDR is competitive to state-of-the-art systems, without using any semi-structured information. Through query composition in "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2104.05883","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/2104.05883/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":"2104.05883","created_at":"2026-07-05T02:31:26.257696+00:00"},{"alias_kind":"arxiv_version","alias_value":"2104.05883v1","created_at":"2026-07-05T02:31:26.257696+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2104.05883","created_at":"2026-07-05T02:31:26.257696+00:00"},{"alias_kind":"pith_short_12","alias_value":"R2YIHYELNB7C","created_at":"2026-07-05T02:31:26.257696+00:00"},{"alias_kind":"pith_short_16","alias_value":"R2YIHYELNB7CYBYO","created_at":"2026-07-05T02:31:26.257696+00:00"},{"alias_kind":"pith_short_8","alias_value":"R2YIHYEL","created_at":"2026-07-05T02:31:26.257696+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.08317","citing_title":"Large Language Models Still Face Challenges in Multi-Hop Reasoning with External Knowledge","ref_index":19,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK","json":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK.json","graph_json":"https://pith.science/api/pith-number/R2YIHYELNB7CYBYO3CZABQLLAK/graph.json","events_json":"https://pith.science/api/pith-number/R2YIHYELNB7CYBYO3CZABQLLAK/events.json","paper":"https://pith.science/paper/R2YIHYEL"},"agent_actions":{"view_html":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK","download_json":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK.json","view_paper":"https://pith.science/paper/R2YIHYEL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2104.05883&json=true","fetch_graph":"https://pith.science/api/pith-number/R2YIHYELNB7CYBYO3CZABQLLAK/graph.json","fetch_events":"https://pith.science/api/pith-number/R2YIHYELNB7CYBYO3CZABQLLAK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK/action/storage_attestation","attest_author":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK/action/author_attestation","sign_citation":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK/action/citation_signature","submit_replication":"https://pith.science/pith/R2YIHYELNB7CYBYO3CZABQLLAK/action/replication_record"}},"created_at":"2026-07-05T02:31:26.257696+00:00","updated_at":"2026-07-05T02:31:26.257696+00:00"}