{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:5QUCAZHLER3HQH77DJ2LSFXZQD","short_pith_number":"pith:5QUCAZHL","canonical_record":{"source":{"id":"2605.13126","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T07:54:20Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"41d69ac569a14dc1e1e86d995b5adf313a60cb1d72b4fe1772a05f10f06b69a5","abstract_canon_sha256":"516421fe25eae985b59b863a4c66099cd8c172e142f17647495a70063562dea4"},"schema_version":"1.0"},"canonical_sha256":"ec282064eb2476781fff1a74b916f980d020d4fd3969cf4c4bd2c01d26e2e263","source":{"kind":"arxiv","id":"2605.13126","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13126","created_at":"2026-05-18T03:08:57Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13126v2","created_at":"2026-05-18T03:08:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13126","created_at":"2026-05-18T03:08:57Z"},{"alias_kind":"pith_short_12","alias_value":"5QUCAZHLER3H","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"5QUCAZHLER3HQH77","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"5QUCAZHL","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:5QUCAZHLER3HQH77DJ2LSFXZQD","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13126","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T07:54:20Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"41d69ac569a14dc1e1e86d995b5adf313a60cb1d72b4fe1772a05f10f06b69a5","abstract_canon_sha256":"516421fe25eae985b59b863a4c66099cd8c172e142f17647495a70063562dea4"},"schema_version":"1.0"},"canonical_sha256":"ec282064eb2476781fff1a74b916f980d020d4fd3969cf4c4bd2c01d26e2e263","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:08:57.888349Z","signature_b64":"V3Z8aDXyQ3GppXFHG6bUaY4yGm4vVv9Mn9++yaG8P+WWQHuiby1BVrYIAbmmxkQYFbEvoSK6vUnPKgJIyP2UAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ec282064eb2476781fff1a74b916f980d020d4fd3969cf4c4bd2c01d26e2e263","last_reissued_at":"2026-05-18T03:08:57.887535Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:08:57.887535Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13126","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":"2026-05-18T03:08:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/0vNPD8yslbUzyPWUg8CmbTSqnBJTW0cGj5J7o8z2Elqvo3VTZPIQiHm1uLwlaYqj6hHtre6LZFPraJDSBvcDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-29T16:03:50.312657Z"},"content_sha256":"6c97098566231396f23013b49f88627c25fb46b2d88d9f8c3bcbc54b551d8105","schema_version":"1.0","event_id":"sha256:6c97098566231396f23013b49f88627c25fb46b2d88d9f8c3bcbc54b551d8105"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:5QUCAZHLER3HQH77DJ2LSFXZQD","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"MLGIB: Multi-Label Graph Information Bottleneck for Expressive and Robust Message Passing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"MLGIB uses constrained information transmission to preserve predictive labels while filtering irrelevant noise during message passing on multi-label graphs.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chaokai Wu, Haofu Shi, Jianghong Ma, Ningxuan Ma, Xiaofeng Zhang","submitted_at":"2026-05-13T07:54:20Z","abstract_excerpt":"Graph Neural Networks (GNNs) suffer from over-squashing in deep message passing, where information from exponentially growing neighborhoods is compressed into fixed-dimensional representations. We show that this issue becomes a distinct failure mode in multi-label graphs: neighboring nodes often share only limited labels while differing across many irrelevant ones, causing predictive signals to be diluted by noisy label information. To address this challenge, we propose the Multi-Label Graph Information Bottleneck (MLGIB), which formulates multi-label message passing as constrained information"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MLGIB balances expressiveness and robustness by preserving predictive label signals while suppressing irrelevant noise through constrained information transmission under a Markovian dependence space.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That neighboring nodes share only limited labels while differing across many irrelevant ones, causing predictive signals to be diluted by noisy label information in a way that standard message passing cannot handle.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MLGIB derives variational bounds for multi-label message passing to maximize predictive information while constraining redundant noise from irrelevant labels.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MLGIB uses constrained information transmission to preserve predictive labels while filtering irrelevant noise during message passing on multi-label graphs.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1b15aad07a4d77fbb3da69640800cfeaeb4bdc0c3a225dc91da9dd6d49864a69"},"source":{"id":"2605.13126","kind":"arxiv","version":2},"verdict":{"id":"227211cb-7ec0-4060-9e66-c54c0ec08d73","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T05:33:25.892527Z","strongest_claim":"MLGIB balances expressiveness and robustness by preserving predictive label signals while suppressing irrelevant noise through constrained information transmission under a Markovian dependence space.","one_line_summary":"MLGIB derives variational bounds for multi-label message passing to maximize predictive information while constraining redundant noise from irrelevant labels.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That neighboring nodes share only limited labels while differing across many irrelevant ones, causing predictive signals to be diluted by noisy label information in a way that standard message passing cannot handle.","pith_extraction_headline":"MLGIB uses constrained information transmission to preserve predictive labels while filtering irrelevant noise during message passing on multi-label graphs."},"references":{"count":42,"sample":[{"doi":"","year":2026,"title":"Niloofar Azizi, Nils M. Kriege, Nicholas J. A. Harvey, and Horst Bischof. Spectral basis learning for expressive graph neural networks in link prediction. InAAAI, pages 19640–19648. AAAI Press, 2026","work_id":"9932cd77-afba-4644-ab67-0237b6915704","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Asela Hevapathige, Asiri Wijesinghe, and Ahad N. Zehmakan. Beyond fixed depth: Adaptive graph neural networks for node classification under varying homophily. InAAAI, pages 21717– 21725. AAAI Press, 2","work_id":"c92dec16-56ab-4751-b6bb-ad7e50a0f881","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"MORGAN: to bridge mixture of experts and spectral graph neural network","work_id":"0bea07e5-b704-42e8-88f5-470f0cd950c6","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Opolka, Pietro Lio, and José Miguel Hernández- Lobato","work_id":"e41c263d-b402-4deb-b684-85e06f117c65","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Towards bridging generalization and expressivity of graph neural networks","work_id":"37477941-3c26-4b8a-a11f-6c35ad7f3055","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":42,"snapshot_sha256":"a6fc1da24093783a46ef8ace10f5fb18ef9b4fa28b0c023ee132d1e635dbe338","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"50721892c4ea5fcbe35cb9d22419d0f397b3d47a935982d7ffeb720e28fde6ba"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"227211cb-7ec0-4060-9e66-c54c0ec08d73"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T03:08:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"yFH382qel6wVPX8tI4o3dw/0ir09o3JhuariYyatNbBI06ospTw8PvsfQGNLbIvhBNrj+GB36/iBku9QYa6rCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-29T16:03:50.313412Z"},"content_sha256":"f850dbc5cb40181e3380dc4d536e8a6b44099d02c9c83945d8ccd27ebf772b40","schema_version":"1.0","event_id":"sha256:f850dbc5cb40181e3380dc4d536e8a6b44099d02c9c83945d8ccd27ebf772b40"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/5QUCAZHLER3HQH77DJ2LSFXZQD/bundle.json","state_url":"https://pith.science/pith/5QUCAZHLER3HQH77DJ2LSFXZQD/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/5QUCAZHLER3HQH77DJ2LSFXZQD/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":"2026-05-29T16:03:50Z","links":{"resolver":"https://pith.science/pith/5QUCAZHLER3HQH77DJ2LSFXZQD","bundle":"https://pith.science/pith/5QUCAZHLER3HQH77DJ2LSFXZQD/bundle.json","state":"https://pith.science/pith/5QUCAZHLER3HQH77DJ2LSFXZQD/state.json","well_known_bundle":"https://pith.science/.well-known/pith/5QUCAZHLER3HQH77DJ2LSFXZQD/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:5QUCAZHLER3HQH77DJ2LSFXZQD","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":"516421fe25eae985b59b863a4c66099cd8c172e142f17647495a70063562dea4","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T07:54:20Z","title_canon_sha256":"41d69ac569a14dc1e1e86d995b5adf313a60cb1d72b4fe1772a05f10f06b69a5"},"schema_version":"1.0","source":{"id":"2605.13126","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13126","created_at":"2026-05-18T03:08:57Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13126v2","created_at":"2026-05-18T03:08:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13126","created_at":"2026-05-18T03:08:57Z"},{"alias_kind":"pith_short_12","alias_value":"5QUCAZHLER3H","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"5QUCAZHLER3HQH77","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"5QUCAZHL","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:f850dbc5cb40181e3380dc4d536e8a6b44099d02c9c83945d8ccd27ebf772b40","target":"graph","created_at":"2026-05-18T03:08:57Z","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":"MLGIB balances expressiveness and robustness by preserving predictive label signals while suppressing irrelevant noise through constrained information transmission under a Markovian dependence space."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That neighboring nodes share only limited labels while differing across many irrelevant ones, causing predictive signals to be diluted by noisy label information in a way that standard message passing cannot handle."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"MLGIB derives variational bounds for multi-label message passing to maximize predictive information while constraining redundant noise from irrelevant labels."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"MLGIB uses constrained information transmission to preserve predictive labels while filtering irrelevant noise during message passing on multi-label graphs."}],"snapshot_sha256":"1b15aad07a4d77fbb3da69640800cfeaeb4bdc0c3a225dc91da9dd6d49864a69"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"50721892c4ea5fcbe35cb9d22419d0f397b3d47a935982d7ffeb720e28fde6ba"},"paper":{"abstract_excerpt":"Graph Neural Networks (GNNs) suffer from over-squashing in deep message passing, where information from exponentially growing neighborhoods is compressed into fixed-dimensional representations. We show that this issue becomes a distinct failure mode in multi-label graphs: neighboring nodes often share only limited labels while differing across many irrelevant ones, causing predictive signals to be diluted by noisy label information. To address this challenge, we propose the Multi-Label Graph Information Bottleneck (MLGIB), which formulates multi-label message passing as constrained information","authors_text":"Chaokai Wu, Haofu Shi, Jianghong Ma, Ningxuan Ma, Xiaofeng Zhang","cross_cats":["cs.AI"],"headline":"MLGIB uses constrained information transmission to preserve predictive labels while filtering irrelevant noise during message passing on multi-label graphs.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T07:54:20Z","title":"MLGIB: Multi-Label Graph Information Bottleneck for Expressive and Robust Message Passing"},"references":{"count":42,"internal_anchors":1,"resolved_work":42,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Niloofar Azizi, Nils M. Kriege, Nicholas J. A. Harvey, and Horst Bischof. Spectral basis learning for expressive graph neural networks in link prediction. InAAAI, pages 19640–19648. AAAI Press, 2026","work_id":"9932cd77-afba-4644-ab67-0237b6915704","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Asela Hevapathige, Asiri Wijesinghe, and Ahad N. Zehmakan. Beyond fixed depth: Adaptive graph neural networks for node classification under varying homophily. InAAAI, pages 21717– 21725. AAAI Press, 2","work_id":"c92dec16-56ab-4751-b6bb-ad7e50a0f881","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"MORGAN: to bridge mixture of experts and spectral graph neural network","work_id":"0bea07e5-b704-42e8-88f5-470f0cd950c6","year":2026},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Opolka, Pietro Lio, and José Miguel Hernández- Lobato","work_id":"e41c263d-b402-4deb-b684-85e06f117c65","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Towards bridging generalization and expressivity of graph neural networks","work_id":"37477941-3c26-4b8a-a11f-6c35ad7f3055","year":2025}],"snapshot_sha256":"a6fc1da24093783a46ef8ace10f5fb18ef9b4fa28b0c023ee132d1e635dbe338"},"source":{"id":"2605.13126","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-15T05:33:25.892527Z","id":"227211cb-7ec0-4060-9e66-c54c0ec08d73","model_set":{"reader":"grok-4.3"},"one_line_summary":"MLGIB derives variational bounds for multi-label message passing to maximize predictive information while constraining redundant noise from irrelevant labels.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"MLGIB uses constrained information transmission to preserve predictive labels while filtering irrelevant noise during message passing on multi-label graphs.","strongest_claim":"MLGIB balances expressiveness and robustness by preserving predictive label signals while suppressing irrelevant noise through constrained information transmission under a Markovian dependence space.","weakest_assumption":"That neighboring nodes share only limited labels while differing across many irrelevant ones, causing predictive signals to be diluted by noisy label information in a way that standard message passing cannot handle."}},"verdict_id":"227211cb-7ec0-4060-9e66-c54c0ec08d73"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6c97098566231396f23013b49f88627c25fb46b2d88d9f8c3bcbc54b551d8105","target":"record","created_at":"2026-05-18T03:08:57Z","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":"516421fe25eae985b59b863a4c66099cd8c172e142f17647495a70063562dea4","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T07:54:20Z","title_canon_sha256":"41d69ac569a14dc1e1e86d995b5adf313a60cb1d72b4fe1772a05f10f06b69a5"},"schema_version":"1.0","source":{"id":"2605.13126","kind":"arxiv","version":2}},"canonical_sha256":"ec282064eb2476781fff1a74b916f980d020d4fd3969cf4c4bd2c01d26e2e263","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ec282064eb2476781fff1a74b916f980d020d4fd3969cf4c4bd2c01d26e2e263","first_computed_at":"2026-05-18T03:08:57.887535Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T03:08:57.887535Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"V3Z8aDXyQ3GppXFHG6bUaY4yGm4vVv9Mn9++yaG8P+WWQHuiby1BVrYIAbmmxkQYFbEvoSK6vUnPKgJIyP2UAQ==","signature_status":"signed_v1","signed_at":"2026-05-18T03:08:57.888349Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13126","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6c97098566231396f23013b49f88627c25fb46b2d88d9f8c3bcbc54b551d8105","sha256:f850dbc5cb40181e3380dc4d536e8a6b44099d02c9c83945d8ccd27ebf772b40"],"state_sha256":"c406fdd90d9b982acc9fb52f4d3720fd04106ab0715956cf85ab34fa27ba7851"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dTux+wOxNBsHZcr9T1nGhZMOPPLpoOSFR+VYFrRbuCyzUbB7iV6ldYDfaOecSEjjqequCRvCA45Abd0LlNMhBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-29T16:03:50.316738Z","bundle_sha256":"62c86204a3a8f4dfae980d03ff89120a29b5bf37f71307a843e95839505da2a9"}}