{ "profile": "wrs-published-research-view-1", "edition": "ratings-e039106f30dec177", "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS", "entry": { "id": "edge-et-al-graphrag-community-summarization-2024", "title": "GraphRAG: building community summaries for questions about a whole collection", "summary": "Finding one relevant paragraph is different from asking what themes run through thousands of documents. GraphRAG extracts entities and relationships, groups them into communities and prepares summaries at several levels. A question can then draw on these summaries through a map-and-reduce process that combines partial answers. Edge, Trinh and colleagues demonstrate the approach on large text collections. The contribution is this graph-and-community summarization pipeline for collection-wide questions, separate from earlier retrieval, community-detection and language-model methods.", "boundary": "The evaluation uses generated global questions and model-judged answer preferences, rather than a complete factual answer key or independent reader study. Gains vary by criterion and hierarchy level. Query-stage context savings exclude graph extraction and summary preparation. Entity links and summaries can be wrong. The original eight-author version is retained; later bylines are not retroactively substituted. This method is not the invention of all graph-based retrieval or a guarantee that a fluent summary preserves every source condition.", "area": "Retrieval-augmented generation", "topic": "Artificial intelligence", "year": 2024, "lane": "HUMAN", "result": "ALGORITHM", "package_id": "edge-et-al-graphrag-community-summarization-2024", "target": { "record_id": "urn:uuid:0fb584da-e097-42e2-a2bb-5538e53d9045", "version": 1, "content_hash": "sha256:c86e0b0aff634e4877637239367bdc47937913aa62fdfea3a9382515f566421a" }, "humans": [ "Darren Edge", "Ha Trinh", "Newman Cheng", "Joshua Bradley", "Alex Chao", "Apurva Mody", "Steven Truitt", "Jonathan Larson" ], "authors_complete": true, "human_role_note": "Source-listed human researchers and engineers for the exact method or resource. The evaluated learning system is not an independently attested scientific author. Earlier methods, funding and data-collection roles remain separate. Individual effort shares and unresolved person identities are not inferred.", "agents": [], "related": [], "subresults": [], "scientific_assessment": { "source_review": "SOURCE_CURATED", "scientific_review": "NOT_REVIEWED_BY_WRS", "formal_replay": "NOT_RUN", "provider_attestation": "NONE" }, "sources": [ { "id": "cur20261017-arxiv-2404-16130v1", "title": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1", "date": "2024-04-24" } ], "points": "160000", "published_grade": { "id": "edge-et-al-graphrag-community-summarization-2024", "entry_id": "edge-et-al-graphrag-community-summarization-2024", "band": "A", "modifier": "+", "target": { "record_id": "urn:uuid:0fb584da-e097-42e2-a2bb-5538e53d9045", "version": 1, "content_hash": "sha256:c86e0b0aff634e4877637239367bdc47937913aa62fdfea3a9382515f566421a" }, "what_changed": "Finding one relevant paragraph is different from asking what themes run through thousands of documents. GraphRAG extracts entities and relationships, groups them into communities and prepares summaries at several levels. A question can then draw on these summaries through a map-and-reduce process that combines partial answers. Edge, Trinh and colleagues demonstrate the approach on large text collections. The contribution is this graph-and-community summarization pipeline for collection-wide questions, separate from earlier retrieval, community-detection and language-model methods.", "scope": "Finding one relevant paragraph is different from asking what themes run through thousands of documents. GraphRAG extracts entities and relationships, groups them into communities and prepares summaries at several levels. A question can then draw on these summaries through a map-and-reduce process that combines partial answers. Edge, Trinh and colleagues demonstrate the approach on large text collections. The contribution is this graph-and-community summarization pipeline for collection-wide questions, separate from earlier retrieval, community-detection and language-model methods. The evaluation uses generated global questions and model-judged answer preferences, rather than a complete factual answer key or independent reader study. Gains vary by criterion and hierarchy level. Query-stage context savings exclude graph extraction and summary preparation. Entity links and summaries can be wrong. The original eight-author version is retained; later bylines are not retroactively substituted. This method is not the invention of all graph-based retrieval or a guarantee that a fluent summary preserves every source condition.", "status": "PROVISIONAL_EDITORIAL", "scientific_correctness_certified": false }, "assessment_disposition": { "status": "ASSESSED", "label": "Provisionally assessed", "reason": "A current grade and separate reference score identify this exact contribution.", "target": { "record_id": "urn:uuid:0fb584da-e097-42e2-a2bb-5538e53d9045", "version": 1, "content_hash": "sha256:c86e0b0aff634e4877637239367bdc47937913aa62fdfea3a9382515f566421a" }, "edition": "ratings-e039106f30dec177" } }, "version": { "kind": "PUBLIC_RECORD_VIEW", "previous": null, "published_at": "2026-10-23T00:00:00Z", "publisher": { "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33", "role": "WRS_EDITORIAL_PUBLICATION" }, "record": { "contributors": [ { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Darren Edge", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Ha Trinh", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Newman Cheng", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Joshua Bradley", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Alex Chao", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Apurva Mody", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Steven Truitt", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } }, { "attribution_basis": "SOURCE_CITED", "credit": "historical_scientific", "entity_kind": "historical_person", "name": "Jonathan Larson", "note": "Archive attribution; detailed role assignment remains source-bound", "roles": [ "formulator" ], "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } } ], "evidence": [ { "description": "Bibliographic/stated-result source; not a WRS proof verification", "id": "source-0", "kind": "source_reference", "source": { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } } ], "origin": "HUMAN_LEGACY", "record_profile": "wrs-local-research-notes-0.1", "record_type": "research_note", "rrp_version": "0.1", "scope": { "description": "The evaluation uses generated global questions and model-judged answer preferences, rather than a complete factual answer key or independent reader study. Gains vary by criterion and hierarchy level. Query-stage context savings exclude graph extraction and summary preparation. Entity links and summaries can be wrong. The original eight-author version is retained; later bylines are not retroactively substituted. This method is not the invention of all graph-based retrieval or a guarantee that a fluent summary preserves every source condition.", "domain": "Retrieval-augmented generation" }, "sources": [ { "citation": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1" } ], "statement": "# GraphRAG: building community summaries for questions about a whole collection\n\n## One paragraph or the whole collection?\n\n“When did this event happen?” may need one passage. “What disagreements recur across this collection?” requires an overview. GraphRAG addresses the second kind of question by organizing extracted entities and relationships into communities, then preparing summaries at several levels. A later query can combine relevant community reports rather than reread every document from the beginning.\n\n## Build the map before navigating it\n\nAn entity can be a person, organization or concept mentioned in text. A relationship connects entities according to an extraction model. Community detection groups parts of this graph, and a language model summarizes the corresponding material. The query stage produces partial answers and combines them. The map depends on the extraction and summaries: two similar names may be different people, and a confident summary can omit an exception.\n\n## Inspect what a summary leaves out\n\nCreate an invented collection of four short notes: two describe a tool’s speed, one describes its setup cost and one explains where it fails. A summary containing only “the tool is fast” loses the setup and failure conditions. Ask your agent for a collection-wide answer that cites all four notes and keeps those distinctions. Then compare it with an answer based on only the first speed note. The exercise tests coverage, not a model’s scientific correctness.\n\n## How the paper measures improvement\n\nThe original evaluation compares answers to generated global questions using model judgments on qualities such as completeness and diversity. Those preferences are useful evidence about the tested output, but they are not a full factual answer key or a study of reader understanding. Query savings also exclude preparing the graph and reports. The attainable next step is to trace one overview claim back to its passages and list any missing conditions.\n\n## Preserve disagreement in the overview\n\nA good collection summary can say that sources disagree or apply to different conditions. Keep links to the relevant notes, including the one that describes a failure. Hand over both the overview and a compact claim-to-source table. That lets another person audit a fluent answer without reconstructing the whole collection.\n\n## Scientific scope\n\nThe evaluation uses generated global questions and model-judged answer preferences, rather than a complete factual answer key or independent reader study. Gains vary by criterion and hierarchy level. Query-stage context savings exclude graph extraction and summary preparation. Entity links and summaries can be wrong. The original eight-author version is retained; later bylines are not retroactively substituted. This method is not the invention of all graph-based retrieval or a guarantee that a fluent summary preserves every source condition.\n\n## Sources for this explanation\n\n- [From Local to Global: A Graph RAG Approach to Query-Focused Summarization](https://arxiv.org/abs/2404.16130v1): Original eight-author2024 version: graph extraction, community hierarchy, map/reduce answering and global-question evaluation; preparation costs and LLM-judged preferences distinguished.\n", "summary": "Finding one relevant paragraph is different from asking what themes run through thousands of documents. GraphRAG extracts entities and relationships, groups them into communities and prepares summaries at several levels. A question can then draw on these summaries through a map-and-reduce process that combines partial answers. Edge, Trinh and colleagues demonstrate the approach on large text collections. The contribution is this graph-and-community summarization pipeline for collection-wide questions, separate from earlier retrieval, community-detection and language-model methods.", "title": "GraphRAG: building community summaries for questions about a whole collection" }, "record_id": "urn:uuid:0fb584da-e097-42e2-a2bb-5538e53d9045", "version": 1, "source_ref": { "record_id": "urn:uuid:0fb584da-e097-42e2-a2bb-5538e53d9045", "version": 1, "content_hash": "sha256:c86e0b0aff634e4877637239367bdc47937913aa62fdfea3a9382515f566421a" }, "export_view": "wrs-public-editorial-view-1", "public_view_hash": "sha256:6a577ce6fd1d3222cdb073049855ef3c1a22235eaa30d73bfed515336993ff9b" }, "target": { "record_id": "urn:uuid:0fb584da-e097-42e2-a2bb-5538e53d9045", "version": 1, "content_hash": "sha256:c86e0b0aff634e4877637239367bdc47937913aa62fdfea3a9382515f566421a" }, "published_grade": { "profile": "wrs-published-grade-1", "edition": "2026-10-28.1", "id": "edge-et-al-graphrag-community-summarization-2024", "entry_id": "edge-et-al-graphrag-community-summarization-2024", "revision": 1, "title": "GraphRAG: building community summaries for questions about a whole collection", "context": "Retrieval-augmented generation", "track": "Reusable scientific method, experimental result or historical framework", "band": "A", "modifier": "+", "status": "PROVISIONAL_EDITORIAL", "assessed_on": "2026-10-23", "evidence_cutoff": "2026-10-23", "achievement_state": "SUPPORTED_SCOPED_RESULT", "what_changed": "Finding one relevant paragraph is different from asking what themes run through thousands of documents. GraphRAG extracts entities and relationships, groups them into communities and prepares summaries at several levels. A question can then draw on these summaries through a map-and-reduce process that combines partial answers. Edge, Trinh and colleagues demonstrate the approach on large text collections. The contribution is this graph-and-community summarization pipeline for collection-wide questions, separate from earlier retrieval, community-detection and language-model methods.", "scope": "Finding one relevant paragraph is different from asking what themes run through thousands of documents. GraphRAG extracts entities and relationships, groups them into communities and prepares summaries at several levels. A question can then draw on these summaries through a map-and-reduce process that combines partial answers. Edge, Trinh and colleagues demonstrate the approach on large text collections. The contribution is this graph-and-community summarization pipeline for collection-wide questions, separate from earlier retrieval, community-detection and language-model methods. The evaluation uses generated global questions and model-judged answer preferences, rather than a complete factual answer key or independent reader study. Gains vary by criterion and hierarchy level. Query-stage context savings exclude graph extraction and summary preparation. Entity links and summaries can be wrong. The original eight-author version is retained; later bylines are not retroactively substituted. This method is not the invention of all graph-based retrieval or a guarantee that a fluent summary preserves every source condition.", "evidence_summary": "The linked original works support the scoped statement and contribution attribution. WRS has not independently replayed experiments or reconstructed a full mathematical proof. The evaluation uses generated global questions and model-judged answer preferences, rather than a complete factual answer key or independent reader study. Gains vary by criterion and hierarchy level. Query-stage context savings exclude graph extraction and summary preparation. Entity links and summaries can be wrong. The original eight-author version is retained; later bylines are not retroactively substituted. This method is not the invention of all graph-based retrieval or a guarantee that a fluent summary preserves every source condition.", "references": [ { "title": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1", "locator": "Original scoped contribution, statement/method and evidence limits; source-specific attribution retained." } ], "roles": [ { "who": "The named authors in the research record", "credit": "SCIENCE", "did": "Source-listed human researchers and engineers for the exact method or resource. The evaluated learning system is not an independently attested scientific author. Earlier methods, funding and data-collection roles remain separate. Individual effort shares and unresolved person identities are not inferred." } ], "target": { "record_id": "urn:uuid:0fb584da-e097-42e2-a2bb-5538e53d9045", "version": 1, "content_hash": "sha256:c86e0b0aff634e4877637239367bdc47937913aa62fdfea3a9382515f566421a" }, "scientific_correctness_certified": false, "independent_expert_calibration": false, "method": "PROPRIETARY", "points": "160000" }, "canonical_agent_submission": false, "scientific_verification_awarded": false, "reader_guide": { "id": "edge-et-al-graphrag-community-summarization-2024", "target": { "record_id": "urn:uuid:0fb584da-e097-42e2-a2bb-5538e53d9045", "version": 1, "content_hash": "sha256:c86e0b0aff634e4877637239367bdc47937913aa62fdfea3a9382515f566421a" }, "sections": [ { "title": "One paragraph or the whole collection?", "text": "“When did this event happen?” may need one passage. “What disagreements recur across this collection?” requires an overview. GraphRAG addresses the second kind of question by organizing extracted entities and relationships into communities, then preparing summaries at several levels. A later query can combine relevant community reports rather than reread every document from the beginning." }, { "title": "Build the map before navigating it", "text": "An entity can be a person, organization or concept mentioned in text. A relationship connects entities according to an extraction model. Community detection groups parts of this graph, and a language model summarizes the corresponding material. The query stage produces partial answers and combines them. The map depends on the extraction and summaries: two similar names may be different people, and a confident summary can omit an exception." }, { "title": "Inspect what a summary leaves out", "text": "Create an invented collection of four short notes: two describe a tool’s speed, one describes its setup cost and one explains where it fails. A summary containing only “the tool is fast” loses the setup and failure conditions. Ask your agent for a collection-wide answer that cites all four notes and keeps those distinctions. Then compare it with an answer based on only the first speed note. The exercise tests coverage, not a model’s scientific correctness." }, { "title": "How the paper measures improvement", "text": "The original evaluation compares answers to generated global questions using model judgments on qualities such as completeness and diversity. Those preferences are useful evidence about the tested output, but they are not a full factual answer key or a study of reader understanding. Query savings also exclude preparing the graph and reports. The attainable next step is to trace one overview claim back to its passages and list any missing conditions." }, { "title": "Preserve disagreement in the overview", "text": "A good collection summary can say that sources disagree or apply to different conditions. Keep links to the relevant notes, including the one that describes a failure. Hand over both the overview and a compact claim-to-source table. That lets another person audit a fluent answer without reconstructing the whole collection." } ], "references": [ { "title": "From Local to Global: A Graph RAG Approach to Query-Focused Summarization", "url": "https://arxiv.org/abs/2404.16130v1", "locator": "Original eight-author2024 version: graph extraction, community hierarchy, map/reduce answering and global-question evaluation; preparation costs and LLM-judged preferences distinguished." } ] }, "reading_status": "EXPLANATORY_GUIDE" }