{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:VJCIVL3EQ5O2ZZ7T3XPQSTU6DS","short_pith_number":"pith:VJCIVL3E","schema_version":"1.0","canonical_sha256":"aa448aaf64875dace7f3dddf094e9e1c8d32edeb7d0be38668a2282a3100648b","source":{"kind":"arxiv","id":"2303.06562","version":2},"attestation_state":"computed","paper":{"title":"ContraNorm: A Contrastive Learning Perspective on Oversmoothing and Beyond","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Tianqi Du, Xiaojun Guo, Yifei Wang, Yisen Wang","submitted_at":"2023-03-12T04:04:51Z","abstract_excerpt":"Oversmoothing is a common phenomenon in a wide range of Graph Neural Networks (GNNs) and Transformers, where performance worsens as the number of layers increases. Instead of characterizing oversmoothing from the view of complete collapse in which representations converge to a single point, we dive into a more general perspective of dimensional collapse in which representations lie in a narrow cone. Accordingly, inspired by the effectiveness of contrastive learning in preventing dimensional collapse, we propose a novel normalization layer called ContraNorm. Intuitively, ContraNorm implicitly s"},"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":"2303.06562","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-03-12T04:04:51Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"046fa91300079213859bc24aece06bdb8c508d856ea36c98f8adffa33ff27f82","abstract_canon_sha256":"68e707f8b9d0144ba62affdbb8ac34cb72ccb359c729b1bdf30786830e9659d4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:06:14.934543Z","signature_b64":"nTjA36q3XfjlnpscGTXP5Pn7pHWlx7UyPokMk9CPxVbOYOHexrXH22Yyovo7BTu1GRSiTFdJrM/X31gJq/MMAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aa448aaf64875dace7f3dddf094e9e1c8d32edeb7d0be38668a2282a3100648b","last_reissued_at":"2026-07-05T06:06:14.934110Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:06:14.934110Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ContraNorm: A Contrastive Learning Perspective on Oversmoothing and Beyond","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Tianqi Du, Xiaojun Guo, Yifei Wang, Yisen Wang","submitted_at":"2023-03-12T04:04:51Z","abstract_excerpt":"Oversmoothing is a common phenomenon in a wide range of Graph Neural Networks (GNNs) and Transformers, where performance worsens as the number of layers increases. Instead of characterizing oversmoothing from the view of complete collapse in which representations converge to a single point, we dive into a more general perspective of dimensional collapse in which representations lie in a narrow cone. Accordingly, inspired by the effectiveness of contrastive learning in preventing dimensional collapse, we propose a novel normalization layer called ContraNorm. Intuitively, ContraNorm implicitly s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.06562","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/2303.06562/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":"2303.06562","created_at":"2026-07-05T06:06:14.934169+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.06562v2","created_at":"2026-07-05T06:06:14.934169+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.06562","created_at":"2026-07-05T06:06:14.934169+00:00"},{"alias_kind":"pith_short_12","alias_value":"VJCIVL3EQ5O2","created_at":"2026-07-05T06:06:14.934169+00:00"},{"alias_kind":"pith_short_16","alias_value":"VJCIVL3EQ5O2ZZ7T","created_at":"2026-07-05T06:06:14.934169+00:00"},{"alias_kind":"pith_short_8","alias_value":"VJCIVL3E","created_at":"2026-07-05T06:06:14.934169+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.05032","citing_title":"Identifying and Understanding Cross-Class Features in Adversarial Training","ref_index":15,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS","json":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS.json","graph_json":"https://pith.science/api/pith-number/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/graph.json","events_json":"https://pith.science/api/pith-number/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/events.json","paper":"https://pith.science/paper/VJCIVL3E"},"agent_actions":{"view_html":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS","download_json":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS.json","view_paper":"https://pith.science/paper/VJCIVL3E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.06562&json=true","fetch_graph":"https://pith.science/api/pith-number/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/graph.json","fetch_events":"https://pith.science/api/pith-number/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/action/storage_attestation","attest_author":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/action/author_attestation","sign_citation":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/action/citation_signature","submit_replication":"https://pith.science/pith/VJCIVL3EQ5O2ZZ7T3XPQSTU6DS/action/replication_record"}},"created_at":"2026-07-05T06:06:14.934169+00:00","updated_at":"2026-07-05T06:06:14.934169+00:00"}