{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:RGEAIVPQPGU7PGOZSI4F5HCPCA","short_pith_number":"pith:RGEAIVPQ","schema_version":"1.0","canonical_sha256":"89880455f079a9f799d992385e9c4f1000ecd63e7483a6f084ce3dff073c4440","source":{"kind":"arxiv","id":"2602.19799","version":2},"attestation_state":"computed","paper":{"title":"Path-conditioned training: a principled way to rescale ReLU neural networks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"stat.ML","authors_text":"Arthur Lebeurrier, R\\'emi Gribonval, Titouan Vayer","submitted_at":"2026-02-23T12:55:48Z","abstract_excerpt":"Despite recent algorithmic advances, we still lack principled ways to leverage the well-documented rescaling symmetries in ReLU neural network parameters. While two properly rescaled weights implement the same function, the training dynamics can be dramatically different. To offer a fresh perspective on exploiting this phenomenon, we build on the recent path-lifting framework, which provides a compact factorization of ReLU networks. We introduce a geometrically motivated criterion to rescale neural network parameters which minimization leads to a conditioning strategy that aligns a kernel in t"},"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":"2602.19799","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2026-02-23T12:55:48Z","cross_cats_sorted":["cs.LG","math.OC"],"title_canon_sha256":"3e8f165754e28d2fcb3a5349a9541269853d7954d679dfcc29859ffaf0cc5184","abstract_canon_sha256":"c7848ad2bc023e08b79e302268758d1ac0720621bd83cddce9e524a40d0cbf6b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:08:45.910248Z","signature_b64":"uJOU7CVCC2jXps09RAsPIFLPgbQvU4Lz5ndduqHQBYK7pP+5XFv116CHc/pj+97xiG5B0BI0MMKzbQYDmXrtDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"89880455f079a9f799d992385e9c4f1000ecd63e7483a6f084ce3dff073c4440","last_reissued_at":"2026-06-04T01:08:45.909305Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:08:45.909305Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Path-conditioned training: a principled way to rescale ReLU neural networks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"stat.ML","authors_text":"Arthur Lebeurrier, R\\'emi Gribonval, Titouan Vayer","submitted_at":"2026-02-23T12:55:48Z","abstract_excerpt":"Despite recent algorithmic advances, we still lack principled ways to leverage the well-documented rescaling symmetries in ReLU neural network parameters. While two properly rescaled weights implement the same function, the training dynamics can be dramatically different. To offer a fresh perspective on exploiting this phenomenon, we build on the recent path-lifting framework, which provides a compact factorization of ReLU networks. We introduce a geometrically motivated criterion to rescale neural network parameters which minimization leads to a conditioning strategy that aligns a kernel in t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.19799","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/2602.19799/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":"2602.19799","created_at":"2026-06-04T01:08:45.909407+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.19799v2","created_at":"2026-06-04T01:08:45.909407+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.19799","created_at":"2026-06-04T01:08:45.909407+00:00"},{"alias_kind":"pith_short_12","alias_value":"RGEAIVPQPGU7","created_at":"2026-06-04T01:08:45.909407+00:00"},{"alias_kind":"pith_short_16","alias_value":"RGEAIVPQPGU7PGOZ","created_at":"2026-06-04T01:08:45.909407+00:00"},{"alias_kind":"pith_short_8","alias_value":"RGEAIVPQ","created_at":"2026-06-04T01:08:45.909407+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/RGEAIVPQPGU7PGOZSI4F5HCPCA","json":"https://pith.science/pith/RGEAIVPQPGU7PGOZSI4F5HCPCA.json","graph_json":"https://pith.science/api/pith-number/RGEAIVPQPGU7PGOZSI4F5HCPCA/graph.json","events_json":"https://pith.science/api/pith-number/RGEAIVPQPGU7PGOZSI4F5HCPCA/events.json","paper":"https://pith.science/paper/RGEAIVPQ"},"agent_actions":{"view_html":"https://pith.science/pith/RGEAIVPQPGU7PGOZSI4F5HCPCA","download_json":"https://pith.science/pith/RGEAIVPQPGU7PGOZSI4F5HCPCA.json","view_paper":"https://pith.science/paper/RGEAIVPQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.19799&json=true","fetch_graph":"https://pith.science/api/pith-number/RGEAIVPQPGU7PGOZSI4F5HCPCA/graph.json","fetch_events":"https://pith.science/api/pith-number/RGEAIVPQPGU7PGOZSI4F5HCPCA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RGEAIVPQPGU7PGOZSI4F5HCPCA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RGEAIVPQPGU7PGOZSI4F5HCPCA/action/storage_attestation","attest_author":"https://pith.science/pith/RGEAIVPQPGU7PGOZSI4F5HCPCA/action/author_attestation","sign_citation":"https://pith.science/pith/RGEAIVPQPGU7PGOZSI4F5HCPCA/action/citation_signature","submit_replication":"https://pith.science/pith/RGEAIVPQPGU7PGOZSI4F5HCPCA/action/replication_record"}},"created_at":"2026-06-04T01:08:45.909407+00:00","updated_at":"2026-06-04T01:08:45.909407+00:00"}