{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:G2NP6PHUD6ZXSGM532WJWOXIFG","short_pith_number":"pith:G2NP6PHU","schema_version":"1.0","canonical_sha256":"369aff3cf41fb379199ddeac9b3ae829bd043272ac1d46cf11d39062379ae8d4","source":{"kind":"arxiv","id":"1805.08522","version":5},"attestation_state":"computed","paper":{"title":"Deep learning generalizes because the parameter-function map is biased towards simple functions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.NE"],"primary_cat":"stat.ML","authors_text":"Ard A. Louis, Chico Q. Camargo, Guillermo Valle-P\\'erez","submitted_at":"2018-05-22T11:51:36Z","abstract_excerpt":"Deep neural networks (DNNs) generalize remarkably well without explicit regularization even in the strongly over-parametrized regime where classical learning theory would instead predict that they would severely overfit. While many proposals for some kind of implicit regularization have been made to rationalise this success, there is no consensus for the fundamental reason why DNNs do not strongly overfit. In this paper, we provide a new explanation. By applying a very general probability-complexity bound recently derived from algorithmic information theory (AIT), we argue that the parameter-f"},"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":"1805.08522","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2018-05-22T11:51:36Z","cross_cats_sorted":["cs.AI","cs.LG","cs.NE"],"title_canon_sha256":"d383e8a3c39ec489829901468d6b442064c304955153d69cbde82bb433a3ad7f","abstract_canon_sha256":"dceeebd7446ed63e8e1e1ea3e389906e85866a9bd1de1fd7f899b4246a633634"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:48:06.689175Z","signature_b64":"lFfq9E1Dm7rg5lfCmC+2Uv9UYHXBjjSUwOh/935+7C3iiCPAmdraT+wiDqtdcz8OUaoXNHm04AxQm+oCN00eCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"369aff3cf41fb379199ddeac9b3ae829bd043272ac1d46cf11d39062379ae8d4","last_reissued_at":"2026-05-17T23:48:06.688787Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:48:06.688787Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Deep learning generalizes because the parameter-function map is biased towards simple functions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.NE"],"primary_cat":"stat.ML","authors_text":"Ard A. Louis, Chico Q. Camargo, Guillermo Valle-P\\'erez","submitted_at":"2018-05-22T11:51:36Z","abstract_excerpt":"Deep neural networks (DNNs) generalize remarkably well without explicit regularization even in the strongly over-parametrized regime where classical learning theory would instead predict that they would severely overfit. While many proposals for some kind of implicit regularization have been made to rationalise this success, there is no consensus for the fundamental reason why DNNs do not strongly overfit. In this paper, we provide a new explanation. By applying a very general probability-complexity bound recently derived from algorithmic information theory (AIT), we argue that the parameter-f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.08522","kind":"arxiv","version":5},"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":"1805.08522","created_at":"2026-05-17T23:48:06.688845+00:00"},{"alias_kind":"arxiv_version","alias_value":"1805.08522v5","created_at":"2026-05-17T23:48:06.688845+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.08522","created_at":"2026-05-17T23:48:06.688845+00:00"},{"alias_kind":"pith_short_12","alias_value":"G2NP6PHUD6ZX","created_at":"2026-05-18T12:32:25.280505+00:00"},{"alias_kind":"pith_short_16","alias_value":"G2NP6PHUD6ZXSGM5","created_at":"2026-05-18T12:32:25.280505+00:00"},{"alias_kind":"pith_short_8","alias_value":"G2NP6PHU","created_at":"2026-05-18T12:32:25.280505+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2603.11161","citing_title":"Algorithmic Task Capture, Computational Complexity, and Inductive Bias of Infinite Transformers","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"1906.01820","citing_title":"Risks from Learned Optimization in Advanced Machine Learning Systems","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11170","citing_title":"Unlearning with Asymmetric Sources: Improved Unlearning-Utility Trade-off with Public Data","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05436","citing_title":"Estimating Implicit Regularization in Deep Learning","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07844","citing_title":"Distributional simplicity bias and effective convexity in Energy Based Models","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG","json":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG.json","graph_json":"https://pith.science/api/pith-number/G2NP6PHUD6ZXSGM532WJWOXIFG/graph.json","events_json":"https://pith.science/api/pith-number/G2NP6PHUD6ZXSGM532WJWOXIFG/events.json","paper":"https://pith.science/paper/G2NP6PHU"},"agent_actions":{"view_html":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG","download_json":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG.json","view_paper":"https://pith.science/paper/G2NP6PHU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1805.08522&json=true","fetch_graph":"https://pith.science/api/pith-number/G2NP6PHUD6ZXSGM532WJWOXIFG/graph.json","fetch_events":"https://pith.science/api/pith-number/G2NP6PHUD6ZXSGM532WJWOXIFG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG/action/storage_attestation","attest_author":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG/action/author_attestation","sign_citation":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG/action/citation_signature","submit_replication":"https://pith.science/pith/G2NP6PHUD6ZXSGM532WJWOXIFG/action/replication_record"}},"created_at":"2026-05-17T23:48:06.688845+00:00","updated_at":"2026-05-17T23:48:06.688845+00:00"}