{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:XNBVF7Q4UAJAFX66JHOS74ARBE","short_pith_number":"pith:XNBVF7Q4","schema_version":"1.0","canonical_sha256":"bb4352fe1ca01202dfde49dd2ff011092d0f45fd8d958a6a2d675f661188838c","source":{"kind":"arxiv","id":"2309.04332","version":2},"attestation_state":"computed","paper":{"title":"Graph Neural Networks Use Graphs When They Shouldn't","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Amir Globerson, Ido Amos, Maya Bechler-Speicher, Ran Gilad-Bachrach","submitted_at":"2023-09-08T13:59:18Z","abstract_excerpt":"Predictions over graphs play a crucial role in various domains, including social networks and medicine. Graph Neural Networks (GNNs) have emerged as the dominant approach for learning on graph data. Although a graph-structure is provided as input to the GNN, in some cases the best solution can be obtained by ignoring it. While GNNs have the ability to ignore the graph- structure in such cases, it is not clear that they will. In this work, we show that GNNs actually tend to overfit the given graph-structure. Namely, they use it even when a better solution can be obtained by ignoring it. We anal"},"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":"2309.04332","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-09-08T13:59:18Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"5aacd164e6268330fa63608ba2e8d31b1d4660480a09ba4a38ffccaa6645d3ce","abstract_canon_sha256":"a84d5fbc1347e26fb43c876c98c1de9d083c32e6f1dfd7510ccc832050d6f5af"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:48:48.150119Z","signature_b64":"SPpmCSuiooqrSxw3JxjOU6yOWgd1FflldJ1zPk5Ytp6G09hxZeZstaioI38f2eYPKoImPJ6gE7QcPQbvvSzCBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bb4352fe1ca01202dfde49dd2ff011092d0f45fd8d958a6a2d675f661188838c","last_reissued_at":"2026-07-05T07:48:48.149773Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:48:48.149773Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Graph Neural Networks Use Graphs When They Shouldn't","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Amir Globerson, Ido Amos, Maya Bechler-Speicher, Ran Gilad-Bachrach","submitted_at":"2023-09-08T13:59:18Z","abstract_excerpt":"Predictions over graphs play a crucial role in various domains, including social networks and medicine. Graph Neural Networks (GNNs) have emerged as the dominant approach for learning on graph data. Although a graph-structure is provided as input to the GNN, in some cases the best solution can be obtained by ignoring it. While GNNs have the ability to ignore the graph- structure in such cases, it is not clear that they will. In this work, we show that GNNs actually tend to overfit the given graph-structure. Namely, they use it even when a better solution can be obtained by ignoring it. We anal"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.04332","kind":"arxiv","version":2},"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/2309.04332/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":"2309.04332","created_at":"2026-07-05T07:48:48.149828+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.04332v2","created_at":"2026-07-05T07:48:48.149828+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.04332","created_at":"2026-07-05T07:48:48.149828+00:00"},{"alias_kind":"pith_short_12","alias_value":"XNBVF7Q4UAJA","created_at":"2026-07-05T07:48:48.149828+00:00"},{"alias_kind":"pith_short_16","alias_value":"XNBVF7Q4UAJAFX66","created_at":"2026-07-05T07:48:48.149828+00:00"},{"alias_kind":"pith_short_8","alias_value":"XNBVF7Q4","created_at":"2026-07-05T07:48:48.149828+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2411.17429","citing_title":"Graph Rewiring in GNNs to Mitigate Over-Squashing and Over-Smoothing: A Survey","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE","json":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE.json","graph_json":"https://pith.science/api/pith-number/XNBVF7Q4UAJAFX66JHOS74ARBE/graph.json","events_json":"https://pith.science/api/pith-number/XNBVF7Q4UAJAFX66JHOS74ARBE/events.json","paper":"https://pith.science/paper/XNBVF7Q4"},"agent_actions":{"view_html":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE","download_json":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE.json","view_paper":"https://pith.science/paper/XNBVF7Q4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.04332&json=true","fetch_graph":"https://pith.science/api/pith-number/XNBVF7Q4UAJAFX66JHOS74ARBE/graph.json","fetch_events":"https://pith.science/api/pith-number/XNBVF7Q4UAJAFX66JHOS74ARBE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE/action/storage_attestation","attest_author":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE/action/author_attestation","sign_citation":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE/action/citation_signature","submit_replication":"https://pith.science/pith/XNBVF7Q4UAJAFX66JHOS74ARBE/action/replication_record"}},"created_at":"2026-07-05T07:48:48.149828+00:00","updated_at":"2026-07-05T07:48:48.149828+00:00"}