{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:Z247TXVX2JOJGBLJF5AVHXP3OB","short_pith_number":"pith:Z247TXVX","schema_version":"1.0","canonical_sha256":"ceb9f9deb7d25c9305692f4153ddfb7055d3a88aab54718028628529321f5eff","source":{"kind":"arxiv","id":"2411.03006","version":4},"attestation_state":"computed","paper":{"title":"Neural Networks and (Virtual) Extended Formulations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.DM","cs.LG","math.OC"],"primary_cat":"math.CO","authors_text":"Christoph Hertrich, Georg Loho","submitted_at":"2024-11-05T11:12:11Z","abstract_excerpt":"Neural networks with piecewise linear activation functions, such as rectified linear units (ReLU) or maxout, are among the most fundamental models in modern machine learning. We make a step towards proving lower bounds on the size of such neural networks by linking their representative capabilities to the notion of the extension complexity $\\mathrm{xc}(P)$ of a polytope $P$. This is a well-studied quantity in combinatorial optimization and polyhedral geometry describing the number of inequalities needed to model $P$ as a linear program. We show that $\\mathrm{xc}(P)$ is a lower bound on the siz"},"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":"2411.03006","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.CO","submitted_at":"2024-11-05T11:12:11Z","cross_cats_sorted":["cs.CC","cs.DM","cs.LG","math.OC"],"title_canon_sha256":"c43c0e3614664d07961150ee261d36fb6d4234538ed229541da77eb4251d0839","abstract_canon_sha256":"1bc47038ed29ccc87482cd0ec6eb6a22b2af0a431a8786c9964db2368915d003"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:04:51.508388Z","signature_b64":"OUItr1n5z2GNl08cXmjIkGX95GX74RpyYiEQmAZnvn3ZR7r11BOGDHZWYnqAcsDI7CO0+Lh/lelGGT/mpDgeBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ceb9f9deb7d25c9305692f4153ddfb7055d3a88aab54718028628529321f5eff","last_reissued_at":"2026-05-29T01:04:51.507861Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:04:51.507861Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Neural Networks and (Virtual) Extended Formulations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.DM","cs.LG","math.OC"],"primary_cat":"math.CO","authors_text":"Christoph Hertrich, Georg Loho","submitted_at":"2024-11-05T11:12:11Z","abstract_excerpt":"Neural networks with piecewise linear activation functions, such as rectified linear units (ReLU) or maxout, are among the most fundamental models in modern machine learning. We make a step towards proving lower bounds on the size of such neural networks by linking their representative capabilities to the notion of the extension complexity $\\mathrm{xc}(P)$ of a polytope $P$. This is a well-studied quantity in combinatorial optimization and polyhedral geometry describing the number of inequalities needed to model $P$ as a linear program. We show that $\\mathrm{xc}(P)$ is a lower bound on the siz"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.03006","kind":"arxiv","version":4},"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/2411.03006/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":"2411.03006","created_at":"2026-05-29T01:04:51.507925+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.03006v4","created_at":"2026-05-29T01:04:51.507925+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.03006","created_at":"2026-05-29T01:04:51.507925+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z247TXVX2JOJ","created_at":"2026-05-29T01:04:51.507925+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z247TXVX2JOJGBLJ","created_at":"2026-05-29T01:04:51.507925+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z247TXVX","created_at":"2026-05-29T01:04:51.507925+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2509.22849","citing_title":"Parameterized Hardness of Zonotope Containment and Neural Network Verification","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10219","citing_title":"Parameterized Complexity of Stationarity Testing for Piecewise-Affine Functions and Shallow CNN Losses","ref_index":48,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB","json":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB.json","graph_json":"https://pith.science/api/pith-number/Z247TXVX2JOJGBLJF5AVHXP3OB/graph.json","events_json":"https://pith.science/api/pith-number/Z247TXVX2JOJGBLJF5AVHXP3OB/events.json","paper":"https://pith.science/paper/Z247TXVX"},"agent_actions":{"view_html":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB","download_json":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB.json","view_paper":"https://pith.science/paper/Z247TXVX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.03006&json=true","fetch_graph":"https://pith.science/api/pith-number/Z247TXVX2JOJGBLJF5AVHXP3OB/graph.json","fetch_events":"https://pith.science/api/pith-number/Z247TXVX2JOJGBLJF5AVHXP3OB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB/action/storage_attestation","attest_author":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB/action/author_attestation","sign_citation":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB/action/citation_signature","submit_replication":"https://pith.science/pith/Z247TXVX2JOJGBLJF5AVHXP3OB/action/replication_record"}},"created_at":"2026-05-29T01:04:51.507925+00:00","updated_at":"2026-05-29T01:04:51.507925+00:00"}