{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:RT7MZVCBHT55UVNFCREUJHHFWV","short_pith_number":"pith:RT7MZVCB","schema_version":"1.0","canonical_sha256":"8cfeccd4413cfbda55a51449449ce5b5691505af19bdc909617e4d8b4eea1028","source":{"kind":"arxiv","id":"2207.06300","version":1},"attestation_state":"computed","paper":{"title":"Re2G: Retrieve, Rerank, Generate","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.IR"],"primary_cat":"cs.CL","authors_text":"Alfio Gliozzo, Ankita Rajaram Naik, Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Michael Glass, Pengshan Cai","submitted_at":"2022-07-13T15:51:40Z","abstract_excerpt":"As demonstrated by GPT-3 and T5, transformers grow in capability as parameter spaces become larger and larger. However, for tasks that require a large amount of knowledge, non-parametric memory allows models to grow dramatically with a sub-linear increase in computational cost and GPU memory requirements. Recent models such as RAG and REALM have introduced retrieval into conditional generation. These models incorporate neural initial retrieval from a corpus of passages. We build on this line of research, proposing Re2G, which combines both neural initial retrieval and reranking into a BART-bas"},"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":"2207.06300","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-07-13T15:51:40Z","cross_cats_sorted":["cs.AI","cs.IR"],"title_canon_sha256":"1d47940a8dfc2e6fc391673891b09c6b680eb696d2ebd4f3e97a031f44cc9105","abstract_canon_sha256":"1b02d9fc9103ec0273defac1a54e741c13fd201c511e416dc2d77a02da9b3e04"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:40:03.639170Z","signature_b64":"i3a81UFdmZrTst4POE+k/mJqdUc8upWEd4+w+jiMtSHZ8DHK7ioinKMHsfvLukVVeN4n7b2Kb81R6Ezgy4jBAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8cfeccd4413cfbda55a51449449ce5b5691505af19bdc909617e4d8b4eea1028","last_reissued_at":"2026-07-05T04:40:03.638626Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:40:03.638626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Re2G: Retrieve, Rerank, Generate","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.IR"],"primary_cat":"cs.CL","authors_text":"Alfio Gliozzo, Ankita Rajaram Naik, Gaetano Rossiello, Md Faisal Mahbub Chowdhury, Michael Glass, Pengshan Cai","submitted_at":"2022-07-13T15:51:40Z","abstract_excerpt":"As demonstrated by GPT-3 and T5, transformers grow in capability as parameter spaces become larger and larger. However, for tasks that require a large amount of knowledge, non-parametric memory allows models to grow dramatically with a sub-linear increase in computational cost and GPU memory requirements. Recent models such as RAG and REALM have introduced retrieval into conditional generation. These models incorporate neural initial retrieval from a corpus of passages. We build on this line of research, proposing Re2G, which combines both neural initial retrieval and reranking into a BART-bas"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2207.06300","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/2207.06300/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":"2207.06300","created_at":"2026-07-05T04:40:03.638695+00:00"},{"alias_kind":"arxiv_version","alias_value":"2207.06300v1","created_at":"2026-07-05T04:40:03.638695+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2207.06300","created_at":"2026-07-05T04:40:03.638695+00:00"},{"alias_kind":"pith_short_12","alias_value":"RT7MZVCBHT55","created_at":"2026-07-05T04:40:03.638695+00:00"},{"alias_kind":"pith_short_16","alias_value":"RT7MZVCBHT55UVNF","created_at":"2026-07-05T04:40:03.638695+00:00"},{"alias_kind":"pith_short_8","alias_value":"RT7MZVCB","created_at":"2026-07-05T04:40:03.638695+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":51,"is_internal_anchor":false},{"citing_arxiv_id":"2510.20505","citing_title":"RELOOP: Recursive Retrieval with Multi-Hop Reasoner and Planners for Heterogeneous QA","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2208.03299","citing_title":"Atlas: Few-shot Learning with Retrieval Augmented Language Models","ref_index":191,"is_internal_anchor":false},{"citing_arxiv_id":"2208.03299","citing_title":"Atlas: Few-shot Learning with Retrieval Augmented Language Models","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV","json":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV.json","graph_json":"https://pith.science/api/pith-number/RT7MZVCBHT55UVNFCREUJHHFWV/graph.json","events_json":"https://pith.science/api/pith-number/RT7MZVCBHT55UVNFCREUJHHFWV/events.json","paper":"https://pith.science/paper/RT7MZVCB"},"agent_actions":{"view_html":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV","download_json":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV.json","view_paper":"https://pith.science/paper/RT7MZVCB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2207.06300&json=true","fetch_graph":"https://pith.science/api/pith-number/RT7MZVCBHT55UVNFCREUJHHFWV/graph.json","fetch_events":"https://pith.science/api/pith-number/RT7MZVCBHT55UVNFCREUJHHFWV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV/action/storage_attestation","attest_author":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV/action/author_attestation","sign_citation":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV/action/citation_signature","submit_replication":"https://pith.science/pith/RT7MZVCBHT55UVNFCREUJHHFWV/action/replication_record"}},"created_at":"2026-07-05T04:40:03.638695+00:00","updated_at":"2026-07-05T04:40:03.638695+00:00"}