{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:WUTIOJN3CF7BPSQBGQXLZYJW4W","short_pith_number":"pith:WUTIOJN3","schema_version":"1.0","canonical_sha256":"b5268725bb117e17ca01342ebce136e580d37d7ce6a59ee3136fd5a0515b3d3c","source":{"kind":"arxiv","id":"1712.08968","version":3},"attestation_state":"computed","paper":{"title":"Spurious Local Minima are Common in Two-Layer ReLU Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Itay Safran, Ohad Shamir","submitted_at":"2017-12-24T21:00:10Z","abstract_excerpt":"We consider the optimization problem associated with training simple ReLU neural networks of the form $\\mathbf{x}\\mapsto \\sum_{i=1}^{k}\\max\\{0,\\mathbf{w}_i^\\top \\mathbf{x}\\}$ with respect to the squared loss. We provide a computer-assisted proof that even if the input distribution is standard Gaussian, even if the dimension is arbitrarily large, and even if the target values are generated by such a network, with orthonormal parameter vectors, the problem can still have spurious local minima once $6\\le k\\le 20$. By a concentration of measure argument, this implies that in high input dimensions,"},"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":"1712.08968","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-12-24T21:00:10Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"7fe1cbcbcfdcbb525658cf10d63f1abab25c1f387d66c8607c61a825dbfe81f3","abstract_canon_sha256":"dfc169c16bbca0049224080961c0e27abb39bc202beee74192d1e3251295d8e2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:08:31.258693Z","signature_b64":"DjUiNoLgJQICayQ37jNhZurZXHIkPgB/A0lIpYXeITnkfXHqv+/kIcSzzSG6NrkR9Fu5gKn8Do290Ipq71a5Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b5268725bb117e17ca01342ebce136e580d37d7ce6a59ee3136fd5a0515b3d3c","last_reissued_at":"2026-05-18T00:08:31.258254Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:08:31.258254Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Spurious Local Minima are Common in Two-Layer ReLU Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Itay Safran, Ohad Shamir","submitted_at":"2017-12-24T21:00:10Z","abstract_excerpt":"We consider the optimization problem associated with training simple ReLU neural networks of the form $\\mathbf{x}\\mapsto \\sum_{i=1}^{k}\\max\\{0,\\mathbf{w}_i^\\top \\mathbf{x}\\}$ with respect to the squared loss. We provide a computer-assisted proof that even if the input distribution is standard Gaussian, even if the dimension is arbitrarily large, and even if the target values are generated by such a network, with orthonormal parameter vectors, the problem can still have spurious local minima once $6\\le k\\le 20$. By a concentration of measure argument, this implies that in high input dimensions,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1712.08968","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":""},"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":"1712.08968","created_at":"2026-05-18T00:08:31.258330+00:00"},{"alias_kind":"arxiv_version","alias_value":"1712.08968v3","created_at":"2026-05-18T00:08:31.258330+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1712.08968","created_at":"2026-05-18T00:08:31.258330+00:00"},{"alias_kind":"pith_short_12","alias_value":"WUTIOJN3CF7B","created_at":"2026-05-18T12:31:53.515858+00:00"},{"alias_kind":"pith_short_16","alias_value":"WUTIOJN3CF7BPSQB","created_at":"2026-05-18T12:31:53.515858+00:00"},{"alias_kind":"pith_short_8","alias_value":"WUTIOJN3","created_at":"2026-05-18T12:31:53.515858+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W","json":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W.json","graph_json":"https://pith.science/api/pith-number/WUTIOJN3CF7BPSQBGQXLZYJW4W/graph.json","events_json":"https://pith.science/api/pith-number/WUTIOJN3CF7BPSQBGQXLZYJW4W/events.json","paper":"https://pith.science/paper/WUTIOJN3"},"agent_actions":{"view_html":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W","download_json":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W.json","view_paper":"https://pith.science/paper/WUTIOJN3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1712.08968&json=true","fetch_graph":"https://pith.science/api/pith-number/WUTIOJN3CF7BPSQBGQXLZYJW4W/graph.json","fetch_events":"https://pith.science/api/pith-number/WUTIOJN3CF7BPSQBGQXLZYJW4W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W/action/storage_attestation","attest_author":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W/action/author_attestation","sign_citation":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W/action/citation_signature","submit_replication":"https://pith.science/pith/WUTIOJN3CF7BPSQBGQXLZYJW4W/action/replication_record"}},"created_at":"2026-05-18T00:08:31.258330+00:00","updated_at":"2026-05-18T00:08:31.258330+00:00"}