{"paper":{"title":"CacheRAG: A Semantic Caching System for Retrieval-Augmented Generation in Knowledge Graph Question Answering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"CacheRAG adds semantic caching to turn stateless LLM planners for knowledge graph questions into continual learners.","cross_cats":["cs.CL"],"primary_cat":"cs.DB","authors_text":"Lei Chen, Yushi Sun","submitted_at":"2026-04-28T23:46:47Z","abstract_excerpt":"The integration of Large Language Models (LLMs) with Retrieval-Augmented Generation (RAG) has significantly advanced Knowledge Graph Question Answering (KGQA). However, existing LLM-driven KGQA systems act as stateless planners, generating retrieval plans in isolation without exploiting historical query patterns: analogous to a database system that optimizes every query from scratch without a plan cache. This fundamental design flaw leads to schema hallucinations and limited retrieval coverage. We propose CacheRAG, a systematic cache-augmented architecture for LLM-based KGQA that transforms st"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"CacheRAG significantly outperforms state-of-the-art baselines (e.g., +13.2% accuracy and +17.5% truthfulness on the CRAG dataset).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the three proposed design principles (schema-agnostic ISR interface, MMR-based diversity retrieval, and bounded heuristic expansion) can be realized without introducing new failure modes or requiring extensive manual tuning of cache policies.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"CacheRAG turns stateless LLM planners for KGQA into continual learners via schema-agnostic parsing, diversity-optimized hierarchical caching, and bounded subgraph expansion, yielding up to 13.2% accuracy gains on benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"CacheRAG adds semantic caching to turn stateless LLM planners for knowledge graph questions into continual learners.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"05dbbf56c6a7cdc133da1f8e01aec45738e361af5addb7950dd5126552bba8f7"},"source":{"id":"2604.26176","kind":"arxiv","version":2},"verdict":{"id":"74cba861-951e-4f37-b46c-416e095aee18","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-07T13:44:20.183890Z","strongest_claim":"CacheRAG significantly outperforms state-of-the-art baselines (e.g., +13.2% accuracy and +17.5% truthfulness on the CRAG dataset).","one_line_summary":"CacheRAG turns stateless LLM planners for KGQA into continual learners via schema-agnostic parsing, diversity-optimized hierarchical caching, and bounded subgraph expansion, yielding up to 13.2% accuracy gains on benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the three proposed design principles (schema-agnostic ISR interface, MMR-based diversity retrieval, and bounded heuristic expansion) can be realized without introducing new failure modes or requiring extensive manual tuning of cache policies.","pith_extraction_headline":"CacheRAG adds semantic caching to turn stateless LLM planners for knowledge graph questions into continual learners."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.26176/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-21T03:33:41.849476Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T20:26:07.153365Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"8ccd88ac3ce4e258d15813830b286ce769033665cdc263488d3c43f0e019d048"},"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"}