{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:FVQGH33FQB44HDRYM3GRB6TWBW","short_pith_number":"pith:FVQGH33F","schema_version":"1.0","canonical_sha256":"2d6063ef658079c38e3866cd10fa760d84527a10a0241fca5b3c6f237fcfdebe","source":{"kind":"arxiv","id":"2103.12011","version":2},"attestation_state":"computed","paper":{"title":"Open Domain Question Answering over Tables via Dense Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jonathan Herzig, Julian Martin Eisenschlos, Syrine Krichene, Thomas M\\\"uller","submitted_at":"2021-03-22T17:01:04Z","abstract_excerpt":"Recent advances in open-domain QA have led to strong models based on dense retrieval, but only focused on retrieving textual passages. In this work, we tackle open-domain QA over tables for the first time, and show that retrieval can be improved by a retriever designed to handle tabular context. We present an effective pre-training procedure for our retriever and improve retrieval quality with mined hard negatives. As relevant datasets are missing, we extract a subset of Natural Questions (Kwiatkowski et al., 2019) into a Table QA dataset. We find that our retriever improves retrieval results "},"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":"2103.12011","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-03-22T17:01:04Z","cross_cats_sorted":[],"title_canon_sha256":"3a4790326866da3276f4860dc4ef0a19319f07a69024bd0027733491541f3593","abstract_canon_sha256":"4a441140002d8e2788573fe48bf16ca3c26845a6e3e2881edcca611239771fa6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:47:43.517070Z","signature_b64":"e4y6BdohkUA9MsDT634RUv1Exlv8dXeG1NXva7i8wNFFYJOIDKxpI0QSfHDxBgHeN+NPu8NJUrKwGDF2SxVnBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2d6063ef658079c38e3866cd10fa760d84527a10a0241fca5b3c6f237fcfdebe","last_reissued_at":"2026-07-05T02:47:43.516665Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:47:43.516665Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Open Domain Question Answering over Tables via Dense Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jonathan Herzig, Julian Martin Eisenschlos, Syrine Krichene, Thomas M\\\"uller","submitted_at":"2021-03-22T17:01:04Z","abstract_excerpt":"Recent advances in open-domain QA have led to strong models based on dense retrieval, but only focused on retrieving textual passages. In this work, we tackle open-domain QA over tables for the first time, and show that retrieval can be improved by a retriever designed to handle tabular context. We present an effective pre-training procedure for our retriever and improve retrieval quality with mined hard negatives. As relevant datasets are missing, we extract a subset of Natural Questions (Kwiatkowski et al., 2019) into a Table QA dataset. We find that our retriever improves retrieval results "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.12011","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/2103.12011/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":"2103.12011","created_at":"2026-07-05T02:47:43.516720+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.12011v2","created_at":"2026-07-05T02:47:43.516720+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.12011","created_at":"2026-07-05T02:47:43.516720+00:00"},{"alias_kind":"pith_short_12","alias_value":"FVQGH33FQB44","created_at":"2026-07-05T02:47:43.516720+00:00"},{"alias_kind":"pith_short_16","alias_value":"FVQGH33FQB44HDRY","created_at":"2026-07-05T02:47:43.516720+00:00"},{"alias_kind":"pith_short_8","alias_value":"FVQGH33F","created_at":"2026-07-05T02:47:43.516720+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23997","citing_title":"ChartWalker: Benchmarking the Cross-Chart RAG Task with Hierarchical Knowledge Graphs","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28365","citing_title":"CAMI: Cost-Aware Agent-Guided Multi-Indexing for Semantic Retrieval","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW","json":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW.json","graph_json":"https://pith.science/api/pith-number/FVQGH33FQB44HDRYM3GRB6TWBW/graph.json","events_json":"https://pith.science/api/pith-number/FVQGH33FQB44HDRYM3GRB6TWBW/events.json","paper":"https://pith.science/paper/FVQGH33F"},"agent_actions":{"view_html":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW","download_json":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW.json","view_paper":"https://pith.science/paper/FVQGH33F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.12011&json=true","fetch_graph":"https://pith.science/api/pith-number/FVQGH33FQB44HDRYM3GRB6TWBW/graph.json","fetch_events":"https://pith.science/api/pith-number/FVQGH33FQB44HDRYM3GRB6TWBW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW/action/storage_attestation","attest_author":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW/action/author_attestation","sign_citation":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW/action/citation_signature","submit_replication":"https://pith.science/pith/FVQGH33FQB44HDRYM3GRB6TWBW/action/replication_record"}},"created_at":"2026-07-05T02:47:43.516720+00:00","updated_at":"2026-07-05T02:47:43.516720+00:00"}