{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:PYVWDLPEW3A3XFOSRONIEU52Z5","short_pith_number":"pith:PYVWDLPE","schema_version":"1.0","canonical_sha256":"7e2b61ade4b6c1bb95d28b9a8253bacf70e9a2fece5542a1c93ede1553b37991","source":{"kind":"arxiv","id":"2205.15242","version":4},"attestation_state":"computed","paper":{"title":"Re-parameterizing Your Optimizers rather than Architectures","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Guiguang Ding, Honghao Chen, Jungong Han, Kaiqi Huang, Xiangyu Zhang, Xiaohan Ding","submitted_at":"2022-05-30T16:55:59Z","abstract_excerpt":"The well-designed structures in neural networks reflect the prior knowledge incorporated into the models. However, though different models have various priors, we are used to training them with model-agnostic optimizers such as SGD. In this paper, we propose to incorporate model-specific prior knowledge into optimizers by modifying the gradients according to a set of model-specific hyper-parameters. Such a methodology is referred to as Gradient Re-parameterization, and the optimizers are named RepOptimizers. For the extreme simplicity of model structure, we focus on a VGG-style plain model and"},"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":"2205.15242","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-05-30T16:55:59Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"a401addd672cf914c80a91383d6b5f5543a49739b251684be20dc5191bb512b1","abstract_canon_sha256":"6d4d18df3d815e230dfea02f3c528355bda40ba07e4f5cef75aa8409966efdbd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:40:06.826410Z","signature_b64":"/YS2XtOCYzCrxAh6UVv8+rB8CRoDtGaKA/640lQPYDU7Wt/EkIAUyixtwTGk3JOCsibU3l3S0851xw1yvvt+BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7e2b61ade4b6c1bb95d28b9a8253bacf70e9a2fece5542a1c93ede1553b37991","last_reissued_at":"2026-07-05T05:40:06.825954Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:40:06.825954Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Re-parameterizing Your Optimizers rather than Architectures","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Guiguang Ding, Honghao Chen, Jungong Han, Kaiqi Huang, Xiangyu Zhang, Xiaohan Ding","submitted_at":"2022-05-30T16:55:59Z","abstract_excerpt":"The well-designed structures in neural networks reflect the prior knowledge incorporated into the models. However, though different models have various priors, we are used to training them with model-agnostic optimizers such as SGD. In this paper, we propose to incorporate model-specific prior knowledge into optimizers by modifying the gradients according to a set of model-specific hyper-parameters. Such a methodology is referred to as Gradient Re-parameterization, and the optimizers are named RepOptimizers. For the extreme simplicity of model structure, we focus on a VGG-style plain model and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.15242","kind":"arxiv","version":4},"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/2205.15242/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":"2205.15242","created_at":"2026-07-05T05:40:06.826019+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.15242v4","created_at":"2026-07-05T05:40:06.826019+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.15242","created_at":"2026-07-05T05:40:06.826019+00:00"},{"alias_kind":"pith_short_12","alias_value":"PYVWDLPEW3A3","created_at":"2026-07-05T05:40:06.826019+00:00"},{"alias_kind":"pith_short_16","alias_value":"PYVWDLPEW3A3XFOS","created_at":"2026-07-05T05:40:06.826019+00:00"},{"alias_kind":"pith_short_8","alias_value":"PYVWDLPE","created_at":"2026-07-05T05:40:06.826019+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.13006","citing_title":"What is YOLOv6? A Deep Insight into the Object Detection Model","ref_index":63,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5","json":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5.json","graph_json":"https://pith.science/api/pith-number/PYVWDLPEW3A3XFOSRONIEU52Z5/graph.json","events_json":"https://pith.science/api/pith-number/PYVWDLPEW3A3XFOSRONIEU52Z5/events.json","paper":"https://pith.science/paper/PYVWDLPE"},"agent_actions":{"view_html":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5","download_json":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5.json","view_paper":"https://pith.science/paper/PYVWDLPE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.15242&json=true","fetch_graph":"https://pith.science/api/pith-number/PYVWDLPEW3A3XFOSRONIEU52Z5/graph.json","fetch_events":"https://pith.science/api/pith-number/PYVWDLPEW3A3XFOSRONIEU52Z5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5/action/storage_attestation","attest_author":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5/action/author_attestation","sign_citation":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5/action/citation_signature","submit_replication":"https://pith.science/pith/PYVWDLPEW3A3XFOSRONIEU52Z5/action/replication_record"}},"created_at":"2026-07-05T05:40:06.826019+00:00","updated_at":"2026-07-05T05:40:06.826019+00:00"}