{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:MDWGGBFWYMXQH6MFZVEOX4QQVZ","short_pith_number":"pith:MDWGGBFW","schema_version":"1.0","canonical_sha256":"60ec6304b6c32f03f985cd48ebf210ae66a8d756310c9e37830ede2dc2973569","source":{"kind":"arxiv","id":"2203.15516","version":2},"attestation_state":"computed","paper":{"title":"Rich Feature Construction for the Optimization-Generalization Dilemma","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"David Lopez-Paz, Jianyu Zhang, L\\'eon Bottou","submitted_at":"2022-03-24T20:39:33Z","abstract_excerpt":"There often is a dilemma between ease of optimization and robust out-of-distribution (OoD) generalization. For instance, many OoD methods rely on penalty terms whose optimization is challenging. They are either too strong to optimize reliably or too weak to achieve their goals.\n  We propose to initialize the networks with a rich representation containing a palette of potentially useful features, ready to be used by even simple models. On the one hand, a rich representation provides a good initialization for the optimizer. On the other hand, it also provides an inductive bias that helps OoD gen"},"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":"2203.15516","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-03-24T20:39:33Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"87c1c256be8ed6ecc959ef0557790d983b40955ba8dc1f7383d10e34e9d099a4","abstract_canon_sha256":"5de54ee1005417b36bfa2b4d3e3a43c328adcd3495a4a779f220f3096550f563"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:38:35.425542Z","signature_b64":"uCt6aDHK9+v2jni3pkKEQUXQQdx3QNyiDU/ma8UJ1Z64W6cuUgxtNi2uUwGfzGD6XCR/NhbT8JTR5pp6WiuyBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"60ec6304b6c32f03f985cd48ebf210ae66a8d756310c9e37830ede2dc2973569","last_reissued_at":"2026-07-05T04:38:35.425138Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:38:35.425138Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rich Feature Construction for the Optimization-Generalization Dilemma","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"David Lopez-Paz, Jianyu Zhang, L\\'eon Bottou","submitted_at":"2022-03-24T20:39:33Z","abstract_excerpt":"There often is a dilemma between ease of optimization and robust out-of-distribution (OoD) generalization. For instance, many OoD methods rely on penalty terms whose optimization is challenging. They are either too strong to optimize reliably or too weak to achieve their goals.\n  We propose to initialize the networks with a rich representation containing a palette of potentially useful features, ready to be used by even simple models. On the one hand, a rich representation provides a good initialization for the optimizer. On the other hand, it also provides an inductive bias that helps OoD gen"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.15516","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/2203.15516/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":"2203.15516","created_at":"2026-07-05T04:38:35.425194+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.15516v2","created_at":"2026-07-05T04:38:35.425194+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.15516","created_at":"2026-07-05T04:38:35.425194+00:00"},{"alias_kind":"pith_short_12","alias_value":"MDWGGBFWYMXQ","created_at":"2026-07-05T04:38:35.425194+00:00"},{"alias_kind":"pith_short_16","alias_value":"MDWGGBFWYMXQH6MF","created_at":"2026-07-05T04:38:35.425194+00:00"},{"alias_kind":"pith_short_8","alias_value":"MDWGGBFW","created_at":"2026-07-05T04:38:35.425194+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.12226","citing_title":"Learning Causality for Modern Machine Learning","ref_index":73,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ","json":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ.json","graph_json":"https://pith.science/api/pith-number/MDWGGBFWYMXQH6MFZVEOX4QQVZ/graph.json","events_json":"https://pith.science/api/pith-number/MDWGGBFWYMXQH6MFZVEOX4QQVZ/events.json","paper":"https://pith.science/paper/MDWGGBFW"},"agent_actions":{"view_html":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ","download_json":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ.json","view_paper":"https://pith.science/paper/MDWGGBFW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.15516&json=true","fetch_graph":"https://pith.science/api/pith-number/MDWGGBFWYMXQH6MFZVEOX4QQVZ/graph.json","fetch_events":"https://pith.science/api/pith-number/MDWGGBFWYMXQH6MFZVEOX4QQVZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ/action/storage_attestation","attest_author":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ/action/author_attestation","sign_citation":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ/action/citation_signature","submit_replication":"https://pith.science/pith/MDWGGBFWYMXQH6MFZVEOX4QQVZ/action/replication_record"}},"created_at":"2026-07-05T04:38:35.425194+00:00","updated_at":"2026-07-05T04:38:35.425194+00:00"}