{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:M6H5WMIDZNUDOYTVRUFAVGWTHX","short_pith_number":"pith:M6H5WMID","schema_version":"1.0","canonical_sha256":"678fdb3103cb683762758d0a0a9ad33df171c14cc800f633890219a2e38b5cda","source":{"kind":"arxiv","id":"2410.18938","version":1},"attestation_state":"computed","paper":{"title":"A Random Matrix Theory Perspective on the Spectrum of Learned Features and Asymptotic Generalization Capabilities","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.ST","stat.TH"],"primary_cat":"stat.ML","authors_text":"Bruno Loureiro, Florent Krzakala, Hugo Cui, Luca Pesce, Yatin Dandi, Yue M. Lu","submitted_at":"2024-10-24T17:24:34Z","abstract_excerpt":"A key property of neural networks is their capacity of adapting to data during training. Yet, our current mathematical understanding of feature learning and its relationship to generalization remain limited. In this work, we provide a random matrix analysis of how fully-connected two-layer neural networks adapt to the target function after a single, but aggressive, gradient descent step. We rigorously establish the equivalence between the updated features and an isotropic spiked random feature model, in the limit of large batch size. For the latter model, we derive a deterministic equivalent d"},"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":"2410.18938","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2024-10-24T17:24:34Z","cross_cats_sorted":["cs.LG","math.ST","stat.TH"],"title_canon_sha256":"60a6164aec44861876922452ffefaec5fd1328ddb4207df58cd4eb06a9ab888e","abstract_canon_sha256":"ad2b0533f94eeea2de13011939aa01cdd314dad6bfa07f486ed341025ca16451"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:25:22.302871Z","signature_b64":"r1A+WLbzF+b28JtDrSHFC9EVwcAwcmZg4sdUiUx/kYBQAc6gjBdd/9RNVg83nxdKd/NqI9fAwnBGcgUHmzB8Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"678fdb3103cb683762758d0a0a9ad33df171c14cc800f633890219a2e38b5cda","last_reissued_at":"2026-07-05T09:25:22.302392Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:25:22.302392Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Random Matrix Theory Perspective on the Spectrum of Learned Features and Asymptotic Generalization Capabilities","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.ST","stat.TH"],"primary_cat":"stat.ML","authors_text":"Bruno Loureiro, Florent Krzakala, Hugo Cui, Luca Pesce, Yatin Dandi, Yue M. Lu","submitted_at":"2024-10-24T17:24:34Z","abstract_excerpt":"A key property of neural networks is their capacity of adapting to data during training. Yet, our current mathematical understanding of feature learning and its relationship to generalization remain limited. In this work, we provide a random matrix analysis of how fully-connected two-layer neural networks adapt to the target function after a single, but aggressive, gradient descent step. We rigorously establish the equivalence between the updated features and an isotropic spiked random feature model, in the limit of large batch size. For the latter model, we derive a deterministic equivalent d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.18938","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/2410.18938/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":"2410.18938","created_at":"2026-07-05T09:25:22.302451+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.18938v1","created_at":"2026-07-05T09:25:22.302451+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.18938","created_at":"2026-07-05T09:25:22.302451+00:00"},{"alias_kind":"pith_short_12","alias_value":"M6H5WMIDZNUD","created_at":"2026-07-05T09:25:22.302451+00:00"},{"alias_kind":"pith_short_16","alias_value":"M6H5WMIDZNUDOYTV","created_at":"2026-07-05T09:25:22.302451+00:00"},{"alias_kind":"pith_short_8","alias_value":"M6H5WMID","created_at":"2026-07-05T09:25:22.302451+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28486","citing_title":"Spectral phase transitions and trainability in neural network learning dynamics","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2603.10067","citing_title":"HTMuon: Improving Muon via Heavy-Tailed Spectral Correction","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07870","citing_title":"Spectral Dynamics in Deep Networks: Feature Learning, Outlier Escape, and Learning Rate Transfer","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21254","citing_title":"Random Matrix Spectra from Boltzmann-Weighted Lattice Ensembles","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14567","citing_title":"Scaling Laws from Sequential Feature Recovery: A Solvable Hierarchical Model","ref_index":115,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07870","citing_title":"Spectral Dynamics in Deep Networks: Feature Learning, Outlier Escape, and Learning Rate Transfer","ref_index":59,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX","json":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX.json","graph_json":"https://pith.science/api/pith-number/M6H5WMIDZNUDOYTVRUFAVGWTHX/graph.json","events_json":"https://pith.science/api/pith-number/M6H5WMIDZNUDOYTVRUFAVGWTHX/events.json","paper":"https://pith.science/paper/M6H5WMID"},"agent_actions":{"view_html":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX","download_json":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX.json","view_paper":"https://pith.science/paper/M6H5WMID","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.18938&json=true","fetch_graph":"https://pith.science/api/pith-number/M6H5WMIDZNUDOYTVRUFAVGWTHX/graph.json","fetch_events":"https://pith.science/api/pith-number/M6H5WMIDZNUDOYTVRUFAVGWTHX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX/action/storage_attestation","attest_author":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX/action/author_attestation","sign_citation":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX/action/citation_signature","submit_replication":"https://pith.science/pith/M6H5WMIDZNUDOYTVRUFAVGWTHX/action/replication_record"}},"created_at":"2026-07-05T09:25:22.302451+00:00","updated_at":"2026-07-05T09:25:22.302451+00:00"}