{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:PFRVYT7D5A7CM65FFUGFHQONVQ","short_pith_number":"pith:PFRVYT7D","schema_version":"1.0","canonical_sha256":"79635c4fe3e83e267ba52d0c53c1cdac29991e2869f28bdc2ebcc0988d83ecba","source":{"kind":"arxiv","id":"2101.00072","version":1},"attestation_state":"computed","paper":{"title":"Explicit regularization and implicit bias in deep network classifiers trained with the square loss","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Qianli Liao, Tomaso Poggio","submitted_at":"2020-12-31T21:07:56Z","abstract_excerpt":"Deep ReLU networks trained with the square loss have been observed to perform well in classification tasks. We provide here a theoretical justification based on analysis of the associated gradient flow. We show that convergence to a solution with the absolute minimum norm is expected when normalization techniques such as Batch Normalization (BN) or Weight Normalization (WN) are used together with Weight Decay (WD). The main property of the minimizers that bounds their expected error is the norm: we prove that among all the close-to-interpolating solutions, the ones associated with smaller Frob"},"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":"2101.00072","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2020-12-31T21:07:56Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"fa3d9d2a6f1bc1051e009eea55580031ad3ef2d901b93ac87d2cdf5e861e0a1b","abstract_canon_sha256":"8b1287feb6604d776be61220f708df7fc63cb365a7e95aff2fe6b4220582e83d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:04:15.275457Z","signature_b64":"xV3bpbh9fV0XbRrjInSrfIKhYVuUHeS3x/hWkkSThclyXFyWI/rq/NNGnUbuac4iB93LTYeuoMeoTbFlTN7GDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"79635c4fe3e83e267ba52d0c53c1cdac29991e2869f28bdc2ebcc0988d83ecba","last_reissued_at":"2026-07-05T02:04:15.275063Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:04:15.275063Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Explicit regularization and implicit bias in deep network classifiers trained with the square loss","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Qianli Liao, Tomaso Poggio","submitted_at":"2020-12-31T21:07:56Z","abstract_excerpt":"Deep ReLU networks trained with the square loss have been observed to perform well in classification tasks. We provide here a theoretical justification based on analysis of the associated gradient flow. We show that convergence to a solution with the absolute minimum norm is expected when normalization techniques such as Batch Normalization (BN) or Weight Normalization (WN) are used together with Weight Decay (WD). The main property of the minimizers that bounds their expected error is the norm: we prove that among all the close-to-interpolating solutions, the ones associated with smaller Frob"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.00072","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/2101.00072/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":"2101.00072","created_at":"2026-07-05T02:04:15.275121+00:00"},{"alias_kind":"arxiv_version","alias_value":"2101.00072v1","created_at":"2026-07-05T02:04:15.275121+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.00072","created_at":"2026-07-05T02:04:15.275121+00:00"},{"alias_kind":"pith_short_12","alias_value":"PFRVYT7D5A7C","created_at":"2026-07-05T02:04:15.275121+00:00"},{"alias_kind":"pith_short_16","alias_value":"PFRVYT7D5A7CM65F","created_at":"2026-07-05T02:04:15.275121+00:00"},{"alias_kind":"pith_short_8","alias_value":"PFRVYT7D","created_at":"2026-07-05T02:04:15.275121+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.02865","citing_title":"Memory-efficient Continual Learning with Neural Collapse Contrastive","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ","json":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ.json","graph_json":"https://pith.science/api/pith-number/PFRVYT7D5A7CM65FFUGFHQONVQ/graph.json","events_json":"https://pith.science/api/pith-number/PFRVYT7D5A7CM65FFUGFHQONVQ/events.json","paper":"https://pith.science/paper/PFRVYT7D"},"agent_actions":{"view_html":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ","download_json":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ.json","view_paper":"https://pith.science/paper/PFRVYT7D","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2101.00072&json=true","fetch_graph":"https://pith.science/api/pith-number/PFRVYT7D5A7CM65FFUGFHQONVQ/graph.json","fetch_events":"https://pith.science/api/pith-number/PFRVYT7D5A7CM65FFUGFHQONVQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ/action/storage_attestation","attest_author":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ/action/author_attestation","sign_citation":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ/action/citation_signature","submit_replication":"https://pith.science/pith/PFRVYT7D5A7CM65FFUGFHQONVQ/action/replication_record"}},"created_at":"2026-07-05T02:04:15.275121+00:00","updated_at":"2026-07-05T02:04:15.275121+00:00"}