{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:GBTZLUEYSY3CZRTDPONHSTLRES","short_pith_number":"pith:GBTZLUEY","canonical_record":{"source":{"id":"2604.25693","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-04-28T14:21:03Z","cross_cats_sorted":[],"title_canon_sha256":"a8fd9834244e23cf67349955ba76ab5bcb3763cce1a05c9c5867b4d0bfffda64","abstract_canon_sha256":"60322e093c9a9231326c89ba22a13de59b70166afa74456c85b3c106efea44cd"},"schema_version":"1.0"},"canonical_sha256":"306795d09896362cc6637b9a794d7124ad35f1a95dccb59c1b79120f410eb51e","source":{"kind":"arxiv","id":"2604.25693","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.25693","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"arxiv_version","alias_value":"2604.25693v2","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.25693","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"pith_short_12","alias_value":"GBTZLUEYSY3C","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"pith_short_16","alias_value":"GBTZLUEYSY3CZRTD","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"pith_short_8","alias_value":"GBTZLUEY","created_at":"2026-07-21T02:21:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:GBTZLUEYSY3CZRTDPONHSTLRES","target":"record","payload":{"canonical_record":{"source":{"id":"2604.25693","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-04-28T14:21:03Z","cross_cats_sorted":[],"title_canon_sha256":"a8fd9834244e23cf67349955ba76ab5bcb3763cce1a05c9c5867b4d0bfffda64","abstract_canon_sha256":"60322e093c9a9231326c89ba22a13de59b70166afa74456c85b3c106efea44cd"},"schema_version":"1.0"},"canonical_sha256":"306795d09896362cc6637b9a794d7124ad35f1a95dccb59c1b79120f410eb51e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-21T02:21:36.304776Z","signature_b64":"VunfKO21z6jWAdYMpBo2mTdYBUgS1nvoteBzZY2Kiz4LkUOg3DGltmyYCCZzeXphwI2eBBLg8HKtqrndESIVBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"306795d09896362cc6637b9a794d7124ad35f1a95dccb59c1b79120f410eb51e","last_reissued_at":"2026-07-21T02:21:36.303902Z","signature_status":"signed_v1","first_computed_at":"2026-07-21T02:21:36.303902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.25693","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-21T02:21:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9YSXAC3/gfMDQ39EK4E73H+6k5ZY+uatm8gRiItlV3dRH0nrd5rjq+Fx0VSALCSvVsbEvQlNY7mpSw5izs/mDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T05:41:53.779916Z"},"content_sha256":"cba90e65e13a464c530f90b82e5de89ce4cb373cd4796bc6e02047df513b7606","schema_version":"1.0","event_id":"sha256:cba90e65e13a464c530f90b82e5de89ce4cb373cd4796bc6e02047df513b7606"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:GBTZLUEYSY3CZRTDPONHSTLRES","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RADD: Retrieval-Augmented Discrete Diffusion for Multi-Modal Knowledge Graph Completion","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"RADD splits retrieval from reranking with a discrete diffusion denoiser to improve multi-modal knowledge graph completion.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bo Li, Guanglin Niu","submitted_at":"2026-04-28T14:21:03Z","abstract_excerpt":"Most multi-modal knowledge graph completion (MMKGC) models use one embedding scorer to conduct both retrieval over the full entity set and final link prediction. We argue that this coupling is a core bottleneck: global high-recall search and local fine-grained disambiguation require different inductive biases. Therefore, we propose a Retrieval-Augmented Discrete Diffusion (RADD) framework to decouple retrieval and reranking for MMKGC. A relation-aware multimodal knowledge graph embedding (KGE) retriever serves as both global retriever and distillation teacher, while a conditional discrete deno"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on three MMKGC benchmarks show that RADD achieves the best performance and consistent gains over strong unimodal, multimodal, and LLM-based baselines, while ablations further verify the contribution of each component.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the coupling of global retrieval and local decision-making in a single embedding scorer is the core bottleneck for MMKGC performance, and that the proposed retriever-plus-denoiser split with distillation will reliably overcome it without new failure modes.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"RADD decouples retrieval and reranking in multi-modal KGC via a relation-aware KGE retriever and conditional discrete denoiser, reporting state-of-the-art results on three benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"RADD splits retrieval from reranking with a discrete diffusion denoiser to improve multi-modal knowledge graph completion.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"8efd0e675c85a08e503068456a1999981e8e78162c2b0c73f8e1bced5fc200e4"},"source":{"id":"2604.25693","kind":"arxiv","version":2},"verdict":{"id":"761c7291-a471-4bd7-b277-908ed6d01a90","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-07T16:20:22.692242Z","strongest_claim":"Experiments on three MMKGC benchmarks show that RADD achieves the best performance and consistent gains over strong unimodal, multimodal, and LLM-based baselines, while ablations further verify the contribution of each component.","one_line_summary":"RADD decouples retrieval and reranking in multi-modal KGC via a relation-aware KGE retriever and conditional discrete denoiser, reporting state-of-the-art results on three benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the coupling of global retrieval and local decision-making in a single embedding scorer is the core bottleneck for MMKGC performance, and that the proposed retriever-plus-denoiser split with distillation will reliably overcome it without new failure modes.","pith_extraction_headline":"RADD splits retrieval from reranking with a discrete diffusion denoiser to improve multi-modal knowledge graph completion."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.25693/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-21T04:35:25.903361Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T20:52:10.942347Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"2413ae3587b5650bae0126a8c7e4be5ebf65419f7e6660f81148c2b482284056"},"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"},"verdict_id":"761c7291-a471-4bd7-b277-908ed6d01a90"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-21T02:21:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pqyW0ZFobKwKCPo7Z8SDd1CkuEPr1qM7R3GuLqEwvGSk1N7sQBr/6yklzc3DX0AkjEdNT5yTzxhJ8qhbFc2iBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-11T05:41:53.780639Z"},"content_sha256":"57d4f3f392429ded5933936f5051ddd92846ab4bdbfe79abbf48fe948f28e1a2","schema_version":"1.0","event_id":"sha256:57d4f3f392429ded5933936f5051ddd92846ab4bdbfe79abbf48fe948f28e1a2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GBTZLUEYSY3CZRTDPONHSTLRES/bundle.json","state_url":"https://pith.science/pith/GBTZLUEYSY3CZRTDPONHSTLRES/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GBTZLUEYSY3CZRTDPONHSTLRES/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-11T05:41:53Z","links":{"resolver":"https://pith.science/pith/GBTZLUEYSY3CZRTDPONHSTLRES","bundle":"https://pith.science/pith/GBTZLUEYSY3CZRTDPONHSTLRES/bundle.json","state":"https://pith.science/pith/GBTZLUEYSY3CZRTDPONHSTLRES/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GBTZLUEYSY3CZRTDPONHSTLRES/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GBTZLUEYSY3CZRTDPONHSTLRES","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"60322e093c9a9231326c89ba22a13de59b70166afa74456c85b3c106efea44cd","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-04-28T14:21:03Z","title_canon_sha256":"a8fd9834244e23cf67349955ba76ab5bcb3763cce1a05c9c5867b4d0bfffda64"},"schema_version":"1.0","source":{"id":"2604.25693","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.25693","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"arxiv_version","alias_value":"2604.25693v2","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.25693","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"pith_short_12","alias_value":"GBTZLUEYSY3C","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"pith_short_16","alias_value":"GBTZLUEYSY3CZRTD","created_at":"2026-07-21T02:21:36Z"},{"alias_kind":"pith_short_8","alias_value":"GBTZLUEY","created_at":"2026-07-21T02:21:36Z"}],"graph_snapshots":[{"event_id":"sha256:57d4f3f392429ded5933936f5051ddd92846ab4bdbfe79abbf48fe948f28e1a2","target":"graph","created_at":"2026-07-21T02:21:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Experiments on three MMKGC benchmarks show that RADD achieves the best performance and consistent gains over strong unimodal, multimodal, and LLM-based baselines, while ablations further verify the contribution of each component."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the coupling of global retrieval and local decision-making in a single embedding scorer is the core bottleneck for MMKGC performance, and that the proposed retriever-plus-denoiser split with distillation will reliably overcome it without new failure modes."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"RADD decouples retrieval and reranking in multi-modal KGC via a relation-aware KGE retriever and conditional discrete denoiser, reporting state-of-the-art results on three benchmarks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"RADD splits retrieval from reranking with a discrete diffusion denoiser to improve multi-modal knowledge graph completion."}],"snapshot_sha256":"8efd0e675c85a08e503068456a1999981e8e78162c2b0c73f8e1bced5fc200e4"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-21T04:35:25.903361Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T20:52:10.942347Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2604.25693/integrity.json","findings":[],"snapshot_sha256":"2413ae3587b5650bae0126a8c7e4be5ebf65419f7e6660f81148c2b482284056","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Most multi-modal knowledge graph completion (MMKGC) models use one embedding scorer to conduct both retrieval over the full entity set and final link prediction. We argue that this coupling is a core bottleneck: global high-recall search and local fine-grained disambiguation require different inductive biases. Therefore, we propose a Retrieval-Augmented Discrete Diffusion (RADD) framework to decouple retrieval and reranking for MMKGC. A relation-aware multimodal knowledge graph embedding (KGE) retriever serves as both global retriever and distillation teacher, while a conditional discrete deno","authors_text":"Bo Li, Guanglin Niu","cross_cats":[],"headline":"RADD splits retrieval from reranking with a discrete diffusion denoiser to improve multi-modal knowledge graph completion.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-04-28T14:21:03Z","title":"RADD: Retrieval-Augmented Discrete Diffusion for Multi-Modal Knowledge Graph Completion"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.25693","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-07T16:20:22.692242Z","id":"761c7291-a471-4bd7-b277-908ed6d01a90","model_set":{"reader":"grok-4.3"},"one_line_summary":"RADD decouples retrieval and reranking in multi-modal KGC via a relation-aware KGE retriever and conditional discrete denoiser, reporting state-of-the-art results on three benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"RADD splits retrieval from reranking with a discrete diffusion denoiser to improve multi-modal knowledge graph completion.","strongest_claim":"Experiments on three MMKGC benchmarks show that RADD achieves the best performance and consistent gains over strong unimodal, multimodal, and LLM-based baselines, while ablations further verify the contribution of each component.","weakest_assumption":"That the coupling of global retrieval and local decision-making in a single embedding scorer is the core bottleneck for MMKGC performance, and that the proposed retriever-plus-denoiser split with distillation will reliably overcome it without new failure modes."}},"verdict_id":"761c7291-a471-4bd7-b277-908ed6d01a90"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:cba90e65e13a464c530f90b82e5de89ce4cb373cd4796bc6e02047df513b7606","target":"record","created_at":"2026-07-21T02:21:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"60322e093c9a9231326c89ba22a13de59b70166afa74456c85b3c106efea44cd","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-04-28T14:21:03Z","title_canon_sha256":"a8fd9834244e23cf67349955ba76ab5bcb3763cce1a05c9c5867b4d0bfffda64"},"schema_version":"1.0","source":{"id":"2604.25693","kind":"arxiv","version":2}},"canonical_sha256":"306795d09896362cc6637b9a794d7124ad35f1a95dccb59c1b79120f410eb51e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"306795d09896362cc6637b9a794d7124ad35f1a95dccb59c1b79120f410eb51e","first_computed_at":"2026-07-21T02:21:36.303902Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-21T02:21:36.303902Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"VunfKO21z6jWAdYMpBo2mTdYBUgS1nvoteBzZY2Kiz4LkUOg3DGltmyYCCZzeXphwI2eBBLg8HKtqrndESIVBA==","signature_status":"signed_v1","signed_at":"2026-07-21T02:21:36.304776Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.25693","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:cba90e65e13a464c530f90b82e5de89ce4cb373cd4796bc6e02047df513b7606","sha256:57d4f3f392429ded5933936f5051ddd92846ab4bdbfe79abbf48fe948f28e1a2"],"state_sha256":"973b6e2656b747f96cb4e55307c1270599b71ab1868b18a25fdfacaa3084bc64"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/6fFFok44co4Y4Q4BMW9Yu3tAmViXR3TTYyv8erY8ShH43ep4dHfhNCYUKDzxFkuKwmlgqBenLV3VKis+/xkCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-11T05:41:53.787869Z","bundle_sha256":"af16b2e43bdf64760185805fb091201556e3d769c18ea3846d2b2b285b73629d"}}