{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3QL6JRLOIR3BM75KFMEHQTPNYD","short_pith_number":"pith:3QL6JRLO","schema_version":"1.0","canonical_sha256":"dc17e4c56e4476167faa2b08784dedc0dd3662796859d9bd7571e899b97e2a12","source":{"kind":"arxiv","id":"2309.17357","version":3},"attestation_state":"computed","paper":{"title":"Module-wise Training of Neural Networks via the Minimizing Movement Scheme","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Emmanuel de B\\'ezenac, Ibrahim Ayed, Patrick Gallinari, Skander Karkar","submitted_at":"2023-09-29T16:03:25Z","abstract_excerpt":"Greedy layer-wise or module-wise training of neural networks is compelling in constrained and on-device settings where memory is limited, as it circumvents a number of problems of end-to-end back-propagation. However, it suffers from a stagnation problem, whereby early layers overfit and deeper layers stop increasing the test accuracy after a certain depth. We propose to solve this issue by introducing a module-wise regularization inspired by the minimizing movement scheme for gradient flows in distribution space. We call the method TRGL for Transport Regularized Greedy Learning and study it 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":"2309.17357","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-09-29T16:03:25Z","cross_cats_sorted":[],"title_canon_sha256":"ed0bb4f972e5b2a734258b035d4e4b89486b45ab2226627bc78603656e2c231c","abstract_canon_sha256":"82130200bf18ff9e75cc583ecf28a68d91fb0224a3c624ec567f83878f24202e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:57:26.907741Z","signature_b64":"mRf39TdetIepi0c8HTej2hryMdPx+TPhwh4+9i1C9kW/dll/7MuY+JpRMeQ2WF0ePAbwBfKQgFaq6suCJYF4CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dc17e4c56e4476167faa2b08784dedc0dd3662796859d9bd7571e899b97e2a12","last_reissued_at":"2026-07-05T06:57:26.907351Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:57:26.907351Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Module-wise Training of Neural Networks via the Minimizing Movement Scheme","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Emmanuel de B\\'ezenac, Ibrahim Ayed, Patrick Gallinari, Skander Karkar","submitted_at":"2023-09-29T16:03:25Z","abstract_excerpt":"Greedy layer-wise or module-wise training of neural networks is compelling in constrained and on-device settings where memory is limited, as it circumvents a number of problems of end-to-end back-propagation. However, it suffers from a stagnation problem, whereby early layers overfit and deeper layers stop increasing the test accuracy after a certain depth. We propose to solve this issue by introducing a module-wise regularization inspired by the minimizing movement scheme for gradient flows in distribution space. We call the method TRGL for Transport Regularized Greedy Learning and study it t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.17357","kind":"arxiv","version":3},"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/2309.17357/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":"2309.17357","created_at":"2026-07-05T06:57:26.907409+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.17357v3","created_at":"2026-07-05T06:57:26.907409+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.17357","created_at":"2026-07-05T06:57:26.907409+00:00"},{"alias_kind":"pith_short_12","alias_value":"3QL6JRLOIR3B","created_at":"2026-07-05T06:57:26.907409+00:00"},{"alias_kind":"pith_short_16","alias_value":"3QL6JRLOIR3BM75K","created_at":"2026-07-05T06:57:26.907409+00:00"},{"alias_kind":"pith_short_8","alias_value":"3QL6JRLO","created_at":"2026-07-05T06:57:26.907409+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/3QL6JRLOIR3BM75KFMEHQTPNYD","json":"https://pith.science/pith/3QL6JRLOIR3BM75KFMEHQTPNYD.json","graph_json":"https://pith.science/api/pith-number/3QL6JRLOIR3BM75KFMEHQTPNYD/graph.json","events_json":"https://pith.science/api/pith-number/3QL6JRLOIR3BM75KFMEHQTPNYD/events.json","paper":"https://pith.science/paper/3QL6JRLO"},"agent_actions":{"view_html":"https://pith.science/pith/3QL6JRLOIR3BM75KFMEHQTPNYD","download_json":"https://pith.science/pith/3QL6JRLOIR3BM75KFMEHQTPNYD.json","view_paper":"https://pith.science/paper/3QL6JRLO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.17357&json=true","fetch_graph":"https://pith.science/api/pith-number/3QL6JRLOIR3BM75KFMEHQTPNYD/graph.json","fetch_events":"https://pith.science/api/pith-number/3QL6JRLOIR3BM75KFMEHQTPNYD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3QL6JRLOIR3BM75KFMEHQTPNYD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3QL6JRLOIR3BM75KFMEHQTPNYD/action/storage_attestation","attest_author":"https://pith.science/pith/3QL6JRLOIR3BM75KFMEHQTPNYD/action/author_attestation","sign_citation":"https://pith.science/pith/3QL6JRLOIR3BM75KFMEHQTPNYD/action/citation_signature","submit_replication":"https://pith.science/pith/3QL6JRLOIR3BM75KFMEHQTPNYD/action/replication_record"}},"created_at":"2026-07-05T06:57:26.907409+00:00","updated_at":"2026-07-05T06:57:26.907409+00:00"}