{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:ELP4V5O2OXOO3C34VLKX67NFLF","short_pith_number":"pith:ELP4V5O2","schema_version":"1.0","canonical_sha256":"22dfcaf5da75dced8b7caad57f7da5595ff8dd825952b2d407cd4ffedaac03a4","source":{"kind":"arxiv","id":"1802.05296","version":4},"attestation_state":"computed","paper":{"title":"Stronger generalization bounds for deep nets via a compression approach","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Behnam Neyshabur, Rong Ge, Sanjeev Arora, Yi Zhang","submitted_at":"2018-02-14T19:38:07Z","abstract_excerpt":"Deep nets generalize well despite having more parameters than the number of training samples. Recent works try to give an explanation using PAC-Bayes and Margin-based analyses, but do not as yet result in sample complexity bounds better than naive parameter counting. The current paper shows generalization bounds that're orders of magnitude better in practice. These rely upon new succinct reparametrizations of the trained net --- a compression that is explicit and efficient. These yield generalization bounds via a simple compression-based framework introduced here. Our results also provide some"},"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":"1802.05296","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-02-14T19:38:07Z","cross_cats_sorted":[],"title_canon_sha256":"050c2c1163ad8b9f42e25098afd712f14560d32821b1799a6dac5fae0d6d5c60","abstract_canon_sha256":"f2c64066bc6d28356db935f500d88cfcbf724d7cfb4991d36574a596c429d551"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:59:53.194121Z","signature_b64":"/ZVM2xfykO4GylVXy1U5oDJbXALecdlKUN9Og9m9L4r+f2+PTIb1tSByEVfsWrvjMcjzqZkFN5tLRHi1bsCtCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"22dfcaf5da75dced8b7caad57f7da5595ff8dd825952b2d407cd4ffedaac03a4","last_reissued_at":"2026-05-17T23:59:53.193434Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:59:53.193434Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stronger generalization bounds for deep nets via a compression approach","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Behnam Neyshabur, Rong Ge, Sanjeev Arora, Yi Zhang","submitted_at":"2018-02-14T19:38:07Z","abstract_excerpt":"Deep nets generalize well despite having more parameters than the number of training samples. Recent works try to give an explanation using PAC-Bayes and Margin-based analyses, but do not as yet result in sample complexity bounds better than naive parameter counting. The current paper shows generalization bounds that're orders of magnitude better in practice. These rely upon new succinct reparametrizations of the trained net --- a compression that is explicit and efficient. These yield generalization bounds via a simple compression-based framework introduced here. Our results also provide some"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1802.05296","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":""},"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":"1802.05296","created_at":"2026-05-17T23:59:53.193530+00:00"},{"alias_kind":"arxiv_version","alias_value":"1802.05296v4","created_at":"2026-05-17T23:59:53.193530+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1802.05296","created_at":"2026-05-17T23:59:53.193530+00:00"},{"alias_kind":"pith_short_12","alias_value":"ELP4V5O2OXOO","created_at":"2026-05-18T12:32:22.470017+00:00"},{"alias_kind":"pith_short_16","alias_value":"ELP4V5O2OXOO3C34","created_at":"2026-05-18T12:32:22.470017+00:00"},{"alias_kind":"pith_short_8","alias_value":"ELP4V5O2","created_at":"2026-05-18T12:32:22.470017+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.15551","citing_title":"Characterizing Learning in Deep Neural Networks using Tractable Algorithmic Complexity Analysis","ref_index":19,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF","json":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF.json","graph_json":"https://pith.science/api/pith-number/ELP4V5O2OXOO3C34VLKX67NFLF/graph.json","events_json":"https://pith.science/api/pith-number/ELP4V5O2OXOO3C34VLKX67NFLF/events.json","paper":"https://pith.science/paper/ELP4V5O2"},"agent_actions":{"view_html":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF","download_json":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF.json","view_paper":"https://pith.science/paper/ELP4V5O2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1802.05296&json=true","fetch_graph":"https://pith.science/api/pith-number/ELP4V5O2OXOO3C34VLKX67NFLF/graph.json","fetch_events":"https://pith.science/api/pith-number/ELP4V5O2OXOO3C34VLKX67NFLF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF/action/storage_attestation","attest_author":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF/action/author_attestation","sign_citation":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF/action/citation_signature","submit_replication":"https://pith.science/pith/ELP4V5O2OXOO3C34VLKX67NFLF/action/replication_record"}},"created_at":"2026-05-17T23:59:53.193530+00:00","updated_at":"2026-05-17T23:59:53.193530+00:00"}