{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:NLA4ME66NIIVAKVETBKOJWCJO3","short_pith_number":"pith:NLA4ME66","schema_version":"1.0","canonical_sha256":"6ac1c613de6a11502aa49854e4d84976c7229249ba2e3f8b4a7be07a3187f9aa","source":{"kind":"arxiv","id":"2003.01054","version":2},"attestation_state":"computed","paper":{"title":"Double Trouble in Double Descent : Bias and Variance(s) in the Lazy Regime","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cond-mat.dis-nn","stat.ML"],"primary_cat":"cs.LG","authors_text":"Florent Krzakala, Giulio Biroli, Maria Refinetti, St\\'ephane d'Ascoli","submitted_at":"2020-03-02T17:39:31Z","abstract_excerpt":"Deep neural networks can achieve remarkable generalization performances while interpolating the training data perfectly. Rather than the U-curve emblematic of the bias-variance trade-off, their test error often follows a \"double descent\" - a mark of the beneficial role of overparametrization. In this work, we develop a quantitative theory for this phenomenon in the so-called lazy learning regime of neural networks, by considering the problem of learning a high-dimensional function with random features regression. We obtain a precise asymptotic expression for the bias-variance decomposition of "},"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":"2003.01054","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-03-02T17:39:31Z","cross_cats_sorted":["cond-mat.dis-nn","stat.ML"],"title_canon_sha256":"9e407e782857d55d4b2595e7d6a09eac983e09c849a41af78056fa7710d6a5e7","abstract_canon_sha256":"2a1eceea1f0fdc648d00cb57d2a0af4906ff9bf166298f3b4117c85a83afcf0f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:52:30.245147Z","signature_b64":"FsWhsgCfNwySmUJkmFb3PaSaqHVYo+4grZrnvahmOjOwd7avKPqkNw4aYAP/R5g1LxZtGq6KYLuVHjMjs97pBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ac1c613de6a11502aa49854e4d84976c7229249ba2e3f8b4a7be07a3187f9aa","last_reissued_at":"2026-07-05T00:52:30.244728Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:52:30.244728Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Double Trouble in Double Descent : Bias and Variance(s) in the Lazy Regime","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cond-mat.dis-nn","stat.ML"],"primary_cat":"cs.LG","authors_text":"Florent Krzakala, Giulio Biroli, Maria Refinetti, St\\'ephane d'Ascoli","submitted_at":"2020-03-02T17:39:31Z","abstract_excerpt":"Deep neural networks can achieve remarkable generalization performances while interpolating the training data perfectly. Rather than the U-curve emblematic of the bias-variance trade-off, their test error often follows a \"double descent\" - a mark of the beneficial role of overparametrization. In this work, we develop a quantitative theory for this phenomenon in the so-called lazy learning regime of neural networks, by considering the problem of learning a high-dimensional function with random features regression. We obtain a precise asymptotic expression for the bias-variance decomposition of "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2003.01054","kind":"arxiv","version":2},"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/2003.01054/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":"2003.01054","created_at":"2026-07-05T00:52:30.244781+00:00"},{"alias_kind":"arxiv_version","alias_value":"2003.01054v2","created_at":"2026-07-05T00:52:30.244781+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2003.01054","created_at":"2026-07-05T00:52:30.244781+00:00"},{"alias_kind":"pith_short_12","alias_value":"NLA4ME66NIIV","created_at":"2026-07-05T00:52:30.244781+00:00"},{"alias_kind":"pith_short_16","alias_value":"NLA4ME66NIIVAKVE","created_at":"2026-07-05T00:52:30.244781+00:00"},{"alias_kind":"pith_short_8","alias_value":"NLA4ME66","created_at":"2026-07-05T00:52:30.244781+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2205.10487","citing_title":"Scaling Laws and Interpretability of Learning from Repeated Data","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14200","citing_title":"How to Scale Mixture-of-Experts: From muP to the Maximally Scale-Stable Parameterization","ref_index":249,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3","json":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3.json","graph_json":"https://pith.science/api/pith-number/NLA4ME66NIIVAKVETBKOJWCJO3/graph.json","events_json":"https://pith.science/api/pith-number/NLA4ME66NIIVAKVETBKOJWCJO3/events.json","paper":"https://pith.science/paper/NLA4ME66"},"agent_actions":{"view_html":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3","download_json":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3.json","view_paper":"https://pith.science/paper/NLA4ME66","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2003.01054&json=true","fetch_graph":"https://pith.science/api/pith-number/NLA4ME66NIIVAKVETBKOJWCJO3/graph.json","fetch_events":"https://pith.science/api/pith-number/NLA4ME66NIIVAKVETBKOJWCJO3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3/action/storage_attestation","attest_author":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3/action/author_attestation","sign_citation":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3/action/citation_signature","submit_replication":"https://pith.science/pith/NLA4ME66NIIVAKVETBKOJWCJO3/action/replication_record"}},"created_at":"2026-07-05T00:52:30.244781+00:00","updated_at":"2026-07-05T00:52:30.244781+00:00"}