{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:XYC33OEF77YMBYNGD6PBXFRM66","short_pith_number":"pith:XYC33OEF","schema_version":"1.0","canonical_sha256":"be05bdb885fff0c0e1a61f9e1b962cf7928b012dda0cb7dbe22994a17a8fac8f","source":{"kind":"arxiv","id":"2606.02608","version":1},"attestation_state":"computed","paper":{"title":"Pruning Deep Neural Networks via the Marchenko--Pastur Distribution","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Houman Owhad, Leonid Berlyand, Theo Bourdais, Yitzchak Shmalo","submitted_at":"2026-05-23T19:44:02Z","abstract_excerpt":"We study a Marchenko--Pastur (MP) random-matrix approach to pruning deep neural networks with very small post-pruning fine-tuning budgets. The main practical contribution is accuracy retention under short calibration and fine-tuning schedules, rather than a long post-pruning reoptimization pipeline. The theory gives deterministic data-path certificates: if the removed component $R$ has small propagated logit effect $L_s \\| R \\psi_1(s) \\|_\\infty$, pruning decreases an elastic-net objective and preserves samples whose dense margin exceeds twice the perturbation. The zero-budget case gives perfec"},"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":"2606.02608","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-23T19:44:02Z","cross_cats_sorted":[],"title_canon_sha256":"e21b90a4d9b03661d7d2cbf772ec6605a220f8aac9c2d0dc17c5371b947b11a8","abstract_canon_sha256":"f6886e0a8c510993e6597003e6ef7ebc6823d07d14d402973a43df2d6b6f425d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T00:05:04.451592Z","signature_b64":"b2vSdlRIfMsYrxecTJM9In65Qsrdq4QYr1U/OEFt0YjV/q5nArxkNu7yV6pn8vwL7t72Jm8vfOiLX00dwrKgBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"be05bdb885fff0c0e1a61f9e1b962cf7928b012dda0cb7dbe22994a17a8fac8f","last_reissued_at":"2026-06-03T00:05:04.451221Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T00:05:04.451221Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Pruning Deep Neural Networks via the Marchenko--Pastur Distribution","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Houman Owhad, Leonid Berlyand, Theo Bourdais, Yitzchak Shmalo","submitted_at":"2026-05-23T19:44:02Z","abstract_excerpt":"We study a Marchenko--Pastur (MP) random-matrix approach to pruning deep neural networks with very small post-pruning fine-tuning budgets. The main practical contribution is accuracy retention under short calibration and fine-tuning schedules, rather than a long post-pruning reoptimization pipeline. The theory gives deterministic data-path certificates: if the removed component $R$ has small propagated logit effect $L_s \\| R \\psi_1(s) \\|_\\infty$, pruning decreases an elastic-net objective and preserves samples whose dense margin exceeds twice the perturbation. The zero-budget case gives perfec"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02608","kind":"arxiv","version":1},"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/2606.02608/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":"2606.02608","created_at":"2026-06-03T00:05:04.451280+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02608v1","created_at":"2026-06-03T00:05:04.451280+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02608","created_at":"2026-06-03T00:05:04.451280+00:00"},{"alias_kind":"pith_short_12","alias_value":"XYC33OEF77YM","created_at":"2026-06-03T00:05:04.451280+00:00"},{"alias_kind":"pith_short_16","alias_value":"XYC33OEF77YMBYNG","created_at":"2026-06-03T00:05:04.451280+00:00"},{"alias_kind":"pith_short_8","alias_value":"XYC33OEF","created_at":"2026-06-03T00:05:04.451280+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/XYC33OEF77YMBYNGD6PBXFRM66","json":"https://pith.science/pith/XYC33OEF77YMBYNGD6PBXFRM66.json","graph_json":"https://pith.science/api/pith-number/XYC33OEF77YMBYNGD6PBXFRM66/graph.json","events_json":"https://pith.science/api/pith-number/XYC33OEF77YMBYNGD6PBXFRM66/events.json","paper":"https://pith.science/paper/XYC33OEF"},"agent_actions":{"view_html":"https://pith.science/pith/XYC33OEF77YMBYNGD6PBXFRM66","download_json":"https://pith.science/pith/XYC33OEF77YMBYNGD6PBXFRM66.json","view_paper":"https://pith.science/paper/XYC33OEF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02608&json=true","fetch_graph":"https://pith.science/api/pith-number/XYC33OEF77YMBYNGD6PBXFRM66/graph.json","fetch_events":"https://pith.science/api/pith-number/XYC33OEF77YMBYNGD6PBXFRM66/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XYC33OEF77YMBYNGD6PBXFRM66/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XYC33OEF77YMBYNGD6PBXFRM66/action/storage_attestation","attest_author":"https://pith.science/pith/XYC33OEF77YMBYNGD6PBXFRM66/action/author_attestation","sign_citation":"https://pith.science/pith/XYC33OEF77YMBYNGD6PBXFRM66/action/citation_signature","submit_replication":"https://pith.science/pith/XYC33OEF77YMBYNGD6PBXFRM66/action/replication_record"}},"created_at":"2026-06-03T00:05:04.451280+00:00","updated_at":"2026-06-03T00:05:04.451280+00:00"}