{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:4QT4Q6VFRQRG2LMY6ZBPGWILP2","short_pith_number":"pith:4QT4Q6VF","schema_version":"1.0","canonical_sha256":"e427c87aa58c226d2d98f642f3590b7ea609a859999384a026cde9133c328b2d","source":{"kind":"arxiv","id":"2006.05205","version":4},"attestation_state":"computed","paper":{"title":"On the Bottleneck of Graph Neural Networks and its Practical Implications","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Eran Yahav, Uri Alon","submitted_at":"2020-06-09T12:04:50Z","abstract_excerpt":"Since the proposal of the graph neural network (GNN) by Gori et al. (2005) and Scarselli et al. (2008), one of the major problems in training GNNs was their struggle to propagate information between distant nodes in the graph. We propose a new explanation for this problem: GNNs are susceptible to a bottleneck when aggregating messages across a long path. This bottleneck causes the over-squashing of exponentially growing information into fixed-size vectors. As a result, GNNs fail to propagate messages originating from distant nodes and perform poorly when the prediction task depends on long-ran"},"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":"2006.05205","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-09T12:04:50Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"bf6bdda59ebf858991b728753eceba02f4a57263352232a9e096425764a0b36c","abstract_canon_sha256":"f7f20f664e9770f69ae9ab430c94da34d1dc2361dd074c2696441d9e5865bcb9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:21:14.142624Z","signature_b64":"6ffhWrh1nX1neXRCHoz2wj9rAvxFsxcrXpAVWVYIgDgjbzB3U3lOsZtZLp9bcZK2yhskE+a2StR2fhqL2IyaBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e427c87aa58c226d2d98f642f3590b7ea609a859999384a026cde9133c328b2d","last_reissued_at":"2026-07-05T02:21:14.142088Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:21:14.142088Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Bottleneck of Graph Neural Networks and its Practical Implications","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Eran Yahav, Uri Alon","submitted_at":"2020-06-09T12:04:50Z","abstract_excerpt":"Since the proposal of the graph neural network (GNN) by Gori et al. (2005) and Scarselli et al. (2008), one of the major problems in training GNNs was their struggle to propagate information between distant nodes in the graph. We propose a new explanation for this problem: GNNs are susceptible to a bottleneck when aggregating messages across a long path. This bottleneck causes the over-squashing of exponentially growing information into fixed-size vectors. As a result, GNNs fail to propagate messages originating from distant nodes and perform poorly when the prediction task depends on long-ran"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.05205","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/2006.05205/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":"2006.05205","created_at":"2026-07-05T02:21:14.142143+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.05205v4","created_at":"2026-07-05T02:21:14.142143+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.05205","created_at":"2026-07-05T02:21:14.142143+00:00"},{"alias_kind":"pith_short_12","alias_value":"4QT4Q6VFRQRG","created_at":"2026-07-05T02:21:14.142143+00:00"},{"alias_kind":"pith_short_16","alias_value":"4QT4Q6VFRQRG2LMY","created_at":"2026-07-05T02:21:14.142143+00:00"},{"alias_kind":"pith_short_8","alias_value":"4QT4Q6VF","created_at":"2026-07-05T02:21:14.142143+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":23,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05257","citing_title":"GeoFlow: Geo-Aware Modeling of Inter-Area Relationships in Origin-Destination Flow Prediction and Generation","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2606.22429","citing_title":"Enhancing LLMs for Graph Tasks via Graph-aware LoRA Generation","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18726","citing_title":"Graph Grounded Cross Attention Transformer Neural Network for Structurally Constrained Full Event Sequence Generation in Predictive Process Monitoring","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2606.19185","citing_title":"AGDN: Learning to Solve Traveling Salesman Problem with Anisotropic Graph Diffusion Network","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07327","citing_title":"Six Open Questions in Machine-Learned Interatomic Potential Foundation Models","ref_index":119,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00698","citing_title":"Quantum machine learning models for graphs","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03260","citing_title":"EqGINO: Equivariant Geometry-Informed Fourier Neural Operators for 3D PDEs","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13834","citing_title":"Topology-Preserving Neural Operator Learning via Hodge Decomposition","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24390","citing_title":"Learning Laplacian Eigenspace with Mass-Aware Neural Operators on Point Clouds","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20906","citing_title":"MMGNN: Multi-level, multi-color graph neural networks for molecular property prediction","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2502.20769","citing_title":"Information Bottleneck-Guided Heterogeneous Graph Learning for Interpretable Neurodevelopmental Disorder Diagnosis","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2601.01123","citing_title":"Learning from Historical Activations in Graph Neural Networks","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21435","citing_title":"Gaussian Sheaf Neural Networks","ref_index":81,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01304","citing_title":"SR-CGCNN: Shared Recurrent Convolution in Crystal Graph Neural Networks for Materials Property Prediction","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2511.06443","citing_title":"How Wide and How Deep? Mitigating Over-Squashing of GNNs via Channel Capacity Constrained Estimation","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13834","citing_title":"Topology-Preserving Neural Operator Learning via Hodge Decomposition","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2104.13478","citing_title":"Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11176","citing_title":"From raw data to neutrino candidates: a neural-network pipeline for Baikal-GVD","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2604.28161","citing_title":"RopeDreamer: A Kinematic Recurrent State Space Model for Dynamics of Flexible Deformable Linear Objects","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08679","citing_title":"Attention-based graph neural networks: a survey","ref_index":176,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01304","citing_title":"SR-CGCNN: Shared Recurrent Convolution in Crystal Graph Neural Networks for Materials Property Prediction","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11791","citing_title":"A Mechanistic Analysis of Looped Reasoning Language Models","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01484","citing_title":"Evaluating LLMs on Large-Scale Graph Property Estimation via Random Walks","ref_index":84,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2","json":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2.json","graph_json":"https://pith.science/api/pith-number/4QT4Q6VFRQRG2LMY6ZBPGWILP2/graph.json","events_json":"https://pith.science/api/pith-number/4QT4Q6VFRQRG2LMY6ZBPGWILP2/events.json","paper":"https://pith.science/paper/4QT4Q6VF"},"agent_actions":{"view_html":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2","download_json":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2.json","view_paper":"https://pith.science/paper/4QT4Q6VF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.05205&json=true","fetch_graph":"https://pith.science/api/pith-number/4QT4Q6VFRQRG2LMY6ZBPGWILP2/graph.json","fetch_events":"https://pith.science/api/pith-number/4QT4Q6VFRQRG2LMY6ZBPGWILP2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2/action/storage_attestation","attest_author":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2/action/author_attestation","sign_citation":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2/action/citation_signature","submit_replication":"https://pith.science/pith/4QT4Q6VFRQRG2LMY6ZBPGWILP2/action/replication_record"}},"created_at":"2026-07-05T02:21:14.142143+00:00","updated_at":"2026-07-05T02:21:14.142143+00:00"}