{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:M2TSVEMQIXKX2BYL4WBXDNYW6T","short_pith_number":"pith:M2TSVEMQ","schema_version":"1.0","canonical_sha256":"66a72a919045d57d070be58371b716f4f5d3bb988918a30f8efd202b40ebf9af","source":{"kind":"arxiv","id":"1712.02501","version":2},"attestation_state":"computed","paper":{"title":"CNNs are Globally Optimal Given Multi-Layer Support","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.NE"],"primary_cat":"cs.LG","authors_text":"Chen Huang, Chen Kong, Simon Lucey","submitted_at":"2017-12-07T06:06:52Z","abstract_excerpt":"Stochastic Gradient Descent (SGD) is the central workhorse for training modern CNNs. Although giving impressive empirical performance it can be slow to converge. In this paper we explore a novel strategy for training a CNN using an alternation strategy that offers substantial speedups during training. We make the following contributions: (i) replace the ReLU non-linearity within a CNN with positive hard-thresholding, (ii) reinterpret this non-linearity as a binary state vector making the entire CNN linear if the multi-layer support is known, and (iii) demonstrate that under certain conditions "},"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":"1712.02501","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-12-07T06:06:52Z","cross_cats_sorted":["cs.CV","cs.NE"],"title_canon_sha256":"a9bbeaed0dc6b156f9c6cbc481a7385464967eaebc7d5c5b2d1b612395ec9ce5","abstract_canon_sha256":"d3afc22f9b19bc1c1b93964906e50cfd988c33941eb15332dfc751443af15645"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:28:01.514806Z","signature_b64":"l0TRnhMUvp2y9Ze2Rb0k5did1zsnbTs3TAsv2/JddsU9hpkG1be61TRb8NGXPj7/R3vTZgpmjdkMBTER1I/aCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"66a72a919045d57d070be58371b716f4f5d3bb988918a30f8efd202b40ebf9af","last_reissued_at":"2026-05-18T00:28:01.514192Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:28:01.514192Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CNNs are Globally Optimal Given Multi-Layer Support","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.NE"],"primary_cat":"cs.LG","authors_text":"Chen Huang, Chen Kong, Simon Lucey","submitted_at":"2017-12-07T06:06:52Z","abstract_excerpt":"Stochastic Gradient Descent (SGD) is the central workhorse for training modern CNNs. Although giving impressive empirical performance it can be slow to converge. In this paper we explore a novel strategy for training a CNN using an alternation strategy that offers substantial speedups during training. We make the following contributions: (i) replace the ReLU non-linearity within a CNN with positive hard-thresholding, (ii) reinterpret this non-linearity as a binary state vector making the entire CNN linear if the multi-layer support is known, and (iii) demonstrate that under certain conditions "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1712.02501","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":""},"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":"1712.02501","created_at":"2026-05-18T00:28:01.514305+00:00"},{"alias_kind":"arxiv_version","alias_value":"1712.02501v2","created_at":"2026-05-18T00:28:01.514305+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1712.02501","created_at":"2026-05-18T00:28:01.514305+00:00"},{"alias_kind":"pith_short_12","alias_value":"M2TSVEMQIXKX","created_at":"2026-05-18T12:31:28.150371+00:00"},{"alias_kind":"pith_short_16","alias_value":"M2TSVEMQIXKX2BYL","created_at":"2026-05-18T12:31:28.150371+00:00"},{"alias_kind":"pith_short_8","alias_value":"M2TSVEMQ","created_at":"2026-05-18T12:31:28.150371+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/M2TSVEMQIXKX2BYL4WBXDNYW6T","json":"https://pith.science/pith/M2TSVEMQIXKX2BYL4WBXDNYW6T.json","graph_json":"https://pith.science/api/pith-number/M2TSVEMQIXKX2BYL4WBXDNYW6T/graph.json","events_json":"https://pith.science/api/pith-number/M2TSVEMQIXKX2BYL4WBXDNYW6T/events.json","paper":"https://pith.science/paper/M2TSVEMQ"},"agent_actions":{"view_html":"https://pith.science/pith/M2TSVEMQIXKX2BYL4WBXDNYW6T","download_json":"https://pith.science/pith/M2TSVEMQIXKX2BYL4WBXDNYW6T.json","view_paper":"https://pith.science/paper/M2TSVEMQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1712.02501&json=true","fetch_graph":"https://pith.science/api/pith-number/M2TSVEMQIXKX2BYL4WBXDNYW6T/graph.json","fetch_events":"https://pith.science/api/pith-number/M2TSVEMQIXKX2BYL4WBXDNYW6T/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M2TSVEMQIXKX2BYL4WBXDNYW6T/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M2TSVEMQIXKX2BYL4WBXDNYW6T/action/storage_attestation","attest_author":"https://pith.science/pith/M2TSVEMQIXKX2BYL4WBXDNYW6T/action/author_attestation","sign_citation":"https://pith.science/pith/M2TSVEMQIXKX2BYL4WBXDNYW6T/action/citation_signature","submit_replication":"https://pith.science/pith/M2TSVEMQIXKX2BYL4WBXDNYW6T/action/replication_record"}},"created_at":"2026-05-18T00:28:01.514305+00:00","updated_at":"2026-05-18T00:28:01.514305+00:00"}