{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:UGVT3ZBLUF3GAL3F6ARFHJJYNR","short_pith_number":"pith:UGVT3ZBL","schema_version":"1.0","canonical_sha256":"a1ab3de42ba176602f65f02253a5386c52f3a9c4c41ab527a7cf7d365a03e221","source":{"kind":"arxiv","id":"2506.22518","version":1},"attestation_state":"computed","paper":{"title":"Weak-to-Strong GraphRAG: Aligning Weak Retrievers with Large Language Models for Graph-based Retrieval Augmented Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bo Han, Chenxi Liu, Deyu Zou, James Cheng, Mufei Li, Pan Li, Siqi Miao, Yongqiang Chen","submitted_at":"2025-06-26T17:40:23Z","abstract_excerpt":"Graph-based retrieval-augmented generation (RAG) enables large language models (LLMs) to ground responses with structured external knowledge from up-to-date knowledge graphs (KGs) and reduce hallucinations. However, LLMs often rely on a weak retriever in graph-based RAG: I) Due to the lack of ground truth, the retriever is often trained on weak supervision, which often introduces spurious signals to the LLMs. II) Due to the abstraction of graph data, the retrieved knowledge is often presented in unorganized forms. To mitigate the issue, we present Refined Graph-based RAG (ReG) to align weak re"},"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":"2506.22518","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-26T17:40:23Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1099ddaefaa81bb6df3173a61d61d1bc99fb4e9981f57e177fe946438b62af0c","abstract_canon_sha256":"d0f67e38c491cd17b04feac05d3c535d2515197cf8c213ef719d9e3e9c74b5f3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:28:39.925376Z","signature_b64":"XDFhF0ugRl0/xkV7j7reeCSXbhrQs9J/V2ATW8T/M3GcG8NER0fJlyF1v36OEl2YYtY/6jpb8acS2yk9WRfmAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a1ab3de42ba176602f65f02253a5386c52f3a9c4c41ab527a7cf7d365a03e221","last_reissued_at":"2026-07-05T11:28:39.924912Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:28:39.924912Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Weak-to-Strong GraphRAG: Aligning Weak Retrievers with Large Language Models for Graph-based Retrieval Augmented Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bo Han, Chenxi Liu, Deyu Zou, James Cheng, Mufei Li, Pan Li, Siqi Miao, Yongqiang Chen","submitted_at":"2025-06-26T17:40:23Z","abstract_excerpt":"Graph-based retrieval-augmented generation (RAG) enables large language models (LLMs) to ground responses with structured external knowledge from up-to-date knowledge graphs (KGs) and reduce hallucinations. However, LLMs often rely on a weak retriever in graph-based RAG: I) Due to the lack of ground truth, the retriever is often trained on weak supervision, which often introduces spurious signals to the LLMs. II) Due to the abstraction of graph data, the retrieved knowledge is often presented in unorganized forms. To mitigate the issue, we present Refined Graph-based RAG (ReG) to align weak re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.22518","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/2506.22518/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":"2506.22518","created_at":"2026-07-05T11:28:39.924967+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.22518v1","created_at":"2026-07-05T11:28:39.924967+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.22518","created_at":"2026-07-05T11:28:39.924967+00:00"},{"alias_kind":"pith_short_12","alias_value":"UGVT3ZBLUF3G","created_at":"2026-07-05T11:28:39.924967+00:00"},{"alias_kind":"pith_short_16","alias_value":"UGVT3ZBLUF3GAL3F","created_at":"2026-07-05T11:28:39.924967+00:00"},{"alias_kind":"pith_short_8","alias_value":"UGVT3ZBL","created_at":"2026-07-05T11:28:39.924967+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.10791","citing_title":"PathISE: Learning Informative Path Supervision for Knowledge Graph Question Answering","ref_index":46,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR","json":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR.json","graph_json":"https://pith.science/api/pith-number/UGVT3ZBLUF3GAL3F6ARFHJJYNR/graph.json","events_json":"https://pith.science/api/pith-number/UGVT3ZBLUF3GAL3F6ARFHJJYNR/events.json","paper":"https://pith.science/paper/UGVT3ZBL"},"agent_actions":{"view_html":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR","download_json":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR.json","view_paper":"https://pith.science/paper/UGVT3ZBL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.22518&json=true","fetch_graph":"https://pith.science/api/pith-number/UGVT3ZBLUF3GAL3F6ARFHJJYNR/graph.json","fetch_events":"https://pith.science/api/pith-number/UGVT3ZBLUF3GAL3F6ARFHJJYNR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR/action/storage_attestation","attest_author":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR/action/author_attestation","sign_citation":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR/action/citation_signature","submit_replication":"https://pith.science/pith/UGVT3ZBLUF3GAL3F6ARFHJJYNR/action/replication_record"}},"created_at":"2026-07-05T11:28:39.924967+00:00","updated_at":"2026-07-05T11:28:39.924967+00:00"}