{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:DM6646NCCOX2VBF47PYLNULRAG","short_pith_number":"pith:DM6646NC","schema_version":"1.0","canonical_sha256":"1b3dee79a213afaa84bcfbf0b6d171019ba36346df26528a947df6446060754c","source":{"kind":"arxiv","id":"1806.08459","version":3},"attestation_state":"computed","paper":{"title":"Topological properties of the set of functions generated by neural networks of fixed size","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.FA"],"primary_cat":"math.GN","authors_text":"Felix Voigtlaender, Mones Raslan, Philipp Petersen","submitted_at":"2018-06-22T00:30:51Z","abstract_excerpt":"We analyze the topological properties of the set of functions that can be implemented by neural networks of a fixed size. Surprisingly, this set has many undesirable properties. It is highly non-convex, except possibly for a few exotic activation functions. Moreover, the set is not closed with respect to $L^p$-norms, $0 < p < \\infty$, for all practically-used activation functions, and also not closed with respect to the $L^\\infty$-norm for all practically-used activation functions except for the ReLU and the parametric ReLU. Finally, the function that maps a family of weights to the function c"},"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":"1806.08459","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.GN","submitted_at":"2018-06-22T00:30:51Z","cross_cats_sorted":["math.FA"],"title_canon_sha256":"25910798bc420ec248188a5eb617aa7f88783730ecf06a32fa5ad9c2ab2169ee","abstract_canon_sha256":"a123194973cf940c993997c214243ecf0848182469b36034c35744b7dd1d64e8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:35:18.094834Z","signature_b64":"HVOOtJfxXfkFQbsT8gd6tVpQB7vANlZurKP8aee99yKa3xP4Re0WIWNJCD1rAnyZSKm92kbEVnFES/++Jw9uDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1b3dee79a213afaa84bcfbf0b6d171019ba36346df26528a947df6446060754c","last_reissued_at":"2026-07-05T00:35:18.094365Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:35:18.094365Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Topological properties of the set of functions generated by neural networks of fixed size","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.FA"],"primary_cat":"math.GN","authors_text":"Felix Voigtlaender, Mones Raslan, Philipp Petersen","submitted_at":"2018-06-22T00:30:51Z","abstract_excerpt":"We analyze the topological properties of the set of functions that can be implemented by neural networks of a fixed size. Surprisingly, this set has many undesirable properties. It is highly non-convex, except possibly for a few exotic activation functions. Moreover, the set is not closed with respect to $L^p$-norms, $0 < p < \\infty$, for all practically-used activation functions, and also not closed with respect to the $L^\\infty$-norm for all practically-used activation functions except for the ReLU and the parametric ReLU. Finally, the function that maps a family of weights to the function c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1806.08459","kind":"arxiv","version":3},"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/1806.08459/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":"1806.08459","created_at":"2026-07-05T00:35:18.094425+00:00"},{"alias_kind":"arxiv_version","alias_value":"1806.08459v3","created_at":"2026-07-05T00:35:18.094425+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1806.08459","created_at":"2026-07-05T00:35:18.094425+00:00"},{"alias_kind":"pith_short_12","alias_value":"DM6646NCCOX2","created_at":"2026-07-05T00:35:18.094425+00:00"},{"alias_kind":"pith_short_16","alias_value":"DM6646NCCOX2VBF4","created_at":"2026-07-05T00:35:18.094425+00:00"},{"alias_kind":"pith_short_8","alias_value":"DM6646NC","created_at":"2026-07-05T00:35:18.094425+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"1908.10407","citing_title":"An Energy Approach to the Solution of Partial Differential Equations in Computational Mechanics via Machine Learning: Concepts, Implementation and Applications","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG","json":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG.json","graph_json":"https://pith.science/api/pith-number/DM6646NCCOX2VBF47PYLNULRAG/graph.json","events_json":"https://pith.science/api/pith-number/DM6646NCCOX2VBF47PYLNULRAG/events.json","paper":"https://pith.science/paper/DM6646NC"},"agent_actions":{"view_html":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG","download_json":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG.json","view_paper":"https://pith.science/paper/DM6646NC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1806.08459&json=true","fetch_graph":"https://pith.science/api/pith-number/DM6646NCCOX2VBF47PYLNULRAG/graph.json","fetch_events":"https://pith.science/api/pith-number/DM6646NCCOX2VBF47PYLNULRAG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG/action/storage_attestation","attest_author":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG/action/author_attestation","sign_citation":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG/action/citation_signature","submit_replication":"https://pith.science/pith/DM6646NCCOX2VBF47PYLNULRAG/action/replication_record"}},"created_at":"2026-07-05T00:35:18.094425+00:00","updated_at":"2026-07-05T00:35:18.094425+00:00"}