{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:LEDWNKC5ZWUURRPBV7VAV7CSLT","short_pith_number":"pith:LEDWNKC5","schema_version":"1.0","canonical_sha256":"590766a85dcda948c5e1afea0afc525cfafc4b5722cdd40b9281300c630f493e","source":{"kind":"arxiv","id":"2504.20734","version":4},"attestation_state":"computed","paper":{"title":"UniversalRAG: Retrieval-Augmented Generation over Corpora of Diverse Modalities and Granularities","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.IR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Jinheon Baek, Kangsan Kim, Soyeong Jeong, Sung Ju Hwang, Woongyeong Yeo","submitted_at":"2025-04-29T13:18:58Z","abstract_excerpt":"Retrieval-Augmented Generation (RAG) has shown substantial promise in improving factual accuracy by grounding model responses with external knowledge relevant to queries. However, most existing approaches are limited to a text-only corpus, and while recent efforts have extended RAG to other modalities such as images and videos, they typically operate over a single modality-specific corpus. In contrast, real-world queries vary widely in the type of knowledge they require, which a single type of knowledge source cannot address. To address this, we introduce UniversalRAG, an any-to-any RAG framew"},"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":"2504.20734","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-29T13:18:58Z","cross_cats_sorted":["cs.AI","cs.CV","cs.IR","cs.LG"],"title_canon_sha256":"eeda79f445d214463184140de2f0faf5da4516b838425525291e2a580b456720","abstract_canon_sha256":"cebe59254dea00f64d1b3ade6c9334845899f1b83f6e2c9e2f11340ce5edfc54"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:04:08.786000Z","signature_b64":"d3EwbN5N/6T2o2RY5xylpsRYKywze3nkgN+4oHiqNaS/7iAe0+CAJD6B68qbHFPq2E9M5aEWeKQjeLXUhAMYAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"590766a85dcda948c5e1afea0afc525cfafc4b5722cdd40b9281300c630f493e","last_reissued_at":"2026-05-20T00:04:08.785249Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:04:08.785249Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"UniversalRAG: Retrieval-Augmented Generation over Corpora of Diverse Modalities and Granularities","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.IR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Jinheon Baek, Kangsan Kim, Soyeong Jeong, Sung Ju Hwang, Woongyeong Yeo","submitted_at":"2025-04-29T13:18:58Z","abstract_excerpt":"Retrieval-Augmented Generation (RAG) has shown substantial promise in improving factual accuracy by grounding model responses with external knowledge relevant to queries. However, most existing approaches are limited to a text-only corpus, and while recent efforts have extended RAG to other modalities such as images and videos, they typically operate over a single modality-specific corpus. In contrast, real-world queries vary widely in the type of knowledge they require, which a single type of knowledge source cannot address. To address this, we introduce UniversalRAG, an any-to-any RAG framew"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.20734","kind":"arxiv","version":4},"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/2504.20734/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":"2504.20734","created_at":"2026-05-20T00:04:08.785370+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.20734v4","created_at":"2026-05-20T00:04:08.785370+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.20734","created_at":"2026-05-20T00:04:08.785370+00:00"},{"alias_kind":"pith_short_12","alias_value":"LEDWNKC5ZWUU","created_at":"2026-05-20T00:04:08.785370+00:00"},{"alias_kind":"pith_short_16","alias_value":"LEDWNKC5ZWUURRPB","created_at":"2026-05-20T00:04:08.785370+00:00"},{"alias_kind":"pith_short_8","alias_value":"LEDWNKC5","created_at":"2026-05-20T00:04:08.785370+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":4,"sample":[{"citing_arxiv_id":"2605.15710","citing_title":"SMMBench: A Benchmark for Source-Distributed Multimodal Agent Memory","ref_index":31,"is_internal_anchor":true},{"citing_arxiv_id":"2505.22095","citing_title":"Mixture-of-Retrieval Experts for Reasoning-Guided Multimodal Knowledge Exploitation","ref_index":41,"is_internal_anchor":true},{"citing_arxiv_id":"2507.04590","citing_title":"VLM2Vec-V2: Advancing Multimodal Embedding for Videos, Images, and Visual Documents","ref_index":28,"is_internal_anchor":true},{"citing_arxiv_id":"2512.20136","citing_title":"M$^3$KG-RAG: Multi-hop Multimodal Knowledge Graph-enhanced Retrieval-Augmented Generation","ref_index":55,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT","json":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT.json","graph_json":"https://pith.science/api/pith-number/LEDWNKC5ZWUURRPBV7VAV7CSLT/graph.json","events_json":"https://pith.science/api/pith-number/LEDWNKC5ZWUURRPBV7VAV7CSLT/events.json","paper":"https://pith.science/paper/LEDWNKC5"},"agent_actions":{"view_html":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT","download_json":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT.json","view_paper":"https://pith.science/paper/LEDWNKC5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.20734&json=true","fetch_graph":"https://pith.science/api/pith-number/LEDWNKC5ZWUURRPBV7VAV7CSLT/graph.json","fetch_events":"https://pith.science/api/pith-number/LEDWNKC5ZWUURRPBV7VAV7CSLT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT/action/storage_attestation","attest_author":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT/action/author_attestation","sign_citation":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT/action/citation_signature","submit_replication":"https://pith.science/pith/LEDWNKC5ZWUURRPBV7VAV7CSLT/action/replication_record"}},"created_at":"2026-05-20T00:04:08.785370+00:00","updated_at":"2026-05-20T00:04:08.785370+00:00"}