{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:4W7BZQLSU5VNDZFSC36RJPW7KP","short_pith_number":"pith:4W7BZQLS","schema_version":"1.0","canonical_sha256":"e5be1cc172a76ad1e4b216fd14bedf53fdf38356dbf34e9b7edbf7508fc20c4c","source":{"kind":"arxiv","id":"2006.08859","version":1},"attestation_state":"computed","paper":{"title":"Minimum Width for Universal Approximation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Chulhee Yun, Jaeho Lee, Jinwoo Shin, Sejun Park","submitted_at":"2020-06-16T01:24:21Z","abstract_excerpt":"The universal approximation property of width-bounded networks has been studied as a dual of classical universal approximation results on depth-bounded networks. However, the critical width enabling the universal approximation has not been exactly characterized in terms of the input dimension $d_x$ and the output dimension $d_y$. In this work, we provide the first definitive result in this direction for networks using the ReLU activation functions: The minimum width required for the universal approximation of the $L^p$ functions is exactly $\\max\\{d_x+1,d_y\\}$. We also prove that the same concl"},"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":"2006.08859","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-16T01:24:21Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"396dc8f5e3c219b38464ac24338b1612d5895bdb7c72e76e9692024190f6b87e","abstract_canon_sha256":"7086d21cc25baa6c2443177938df1583f86f36d5392dd0e34c5c99cef238aa85"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:10:43.793914Z","signature_b64":"55xCbij/oJNgCdUcpfFsf+ekrsMsPGeV38qRPo/RMm43/fduIV7SgOb4J/3ctaVBZ4S2G+MarlanM5fzbE6pCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e5be1cc172a76ad1e4b216fd14bedf53fdf38356dbf34e9b7edbf7508fc20c4c","last_reissued_at":"2026-07-05T01:10:43.793483Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:10:43.793483Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Minimum Width for Universal Approximation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Chulhee Yun, Jaeho Lee, Jinwoo Shin, Sejun Park","submitted_at":"2020-06-16T01:24:21Z","abstract_excerpt":"The universal approximation property of width-bounded networks has been studied as a dual of classical universal approximation results on depth-bounded networks. However, the critical width enabling the universal approximation has not been exactly characterized in terms of the input dimension $d_x$ and the output dimension $d_y$. In this work, we provide the first definitive result in this direction for networks using the ReLU activation functions: The minimum width required for the universal approximation of the $L^p$ functions is exactly $\\max\\{d_x+1,d_y\\}$. We also prove that the same concl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.08859","kind":"arxiv","version":1},"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/2006.08859/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":"2006.08859","created_at":"2026-07-05T01:10:43.793548+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.08859v1","created_at":"2026-07-05T01:10:43.793548+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.08859","created_at":"2026-07-05T01:10:43.793548+00:00"},{"alias_kind":"pith_short_12","alias_value":"4W7BZQLSU5VN","created_at":"2026-07-05T01:10:43.793548+00:00"},{"alias_kind":"pith_short_16","alias_value":"4W7BZQLSU5VNDZFS","created_at":"2026-07-05T01:10:43.793548+00:00"},{"alias_kind":"pith_short_8","alias_value":"4W7BZQLS","created_at":"2026-07-05T01:10:43.793548+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23364","citing_title":"Convergence of Gradient Descent for General Neural Network Architectures Beyond the NTK Regime","ref_index":116,"is_internal_anchor":false},{"citing_arxiv_id":"2509.21280","citing_title":"Model reduction of parametric ordinary differential equations via autoencoders: representation properties and convergence analysis","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2001.04385","citing_title":"Universal Differential Equations for Scientific Machine Learning","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14521","citing_title":"Enjoy Your Layer Normalization with the Computational Efficiency of RMSNorm","ref_index":75,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP","json":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP.json","graph_json":"https://pith.science/api/pith-number/4W7BZQLSU5VNDZFSC36RJPW7KP/graph.json","events_json":"https://pith.science/api/pith-number/4W7BZQLSU5VNDZFSC36RJPW7KP/events.json","paper":"https://pith.science/paper/4W7BZQLS"},"agent_actions":{"view_html":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP","download_json":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP.json","view_paper":"https://pith.science/paper/4W7BZQLS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.08859&json=true","fetch_graph":"https://pith.science/api/pith-number/4W7BZQLSU5VNDZFSC36RJPW7KP/graph.json","fetch_events":"https://pith.science/api/pith-number/4W7BZQLSU5VNDZFSC36RJPW7KP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP/action/storage_attestation","attest_author":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP/action/author_attestation","sign_citation":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP/action/citation_signature","submit_replication":"https://pith.science/pith/4W7BZQLSU5VNDZFSC36RJPW7KP/action/replication_record"}},"created_at":"2026-07-05T01:10:43.793548+00:00","updated_at":"2026-07-05T01:10:43.793548+00:00"}