{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4VCGMWYD4KBARY5X7BCHWKQ52E","short_pith_number":"pith:4VCGMWYD","schema_version":"1.0","canonical_sha256":"e544665b03e28208e3b7f8447b2a1dd13029fd03853a9c4008ac90295ed3cc60","source":{"kind":"arxiv","id":"2505.15133","version":1},"attestation_state":"computed","paper":{"title":"DeepKD: A Deeply Decoupled and Denoised Knowledge Distillation Trainer","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Haiduo Huang, Jiangcheng Song, Pengju Ren, Yadong Zhang","submitted_at":"2025-05-21T05:38:57Z","abstract_excerpt":"Recent advances in knowledge distillation have emphasized the importance of decoupling different knowledge components. While existing methods utilize momentum mechanisms to separate task-oriented and distillation gradients, they overlook the inherent conflict between target-class and non-target-class knowledge flows. Furthermore, low-confidence dark knowledge in non-target classes introduces noisy signals that hinder effective knowledge transfer. To address these limitations, we propose DeepKD, a novel training framework that integrates dual-level decoupling with adaptive denoising. First, thr"},"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":"2505.15133","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-21T05:38:57Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"8db4a5a51450e136ca30807a14e7d3205f190433b46bc7bc0204f7b18c72f6c0","abstract_canon_sha256":"1e2198f94eb9e2d94091ea01380e69d4b611a85004ed51ce7df3d0f2bb981c6d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:06:41.107394Z","signature_b64":"HgoP/DthQtqx4jME86dRt3mZTY30tVxtwtL3tN/KjZUkno4Zcb5zOTTPcusrBH3t4HTqP08W7w1esGfcA0QuAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e544665b03e28208e3b7f8447b2a1dd13029fd03853a9c4008ac90295ed3cc60","last_reissued_at":"2026-07-05T11:06:41.106931Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:06:41.106931Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DeepKD: A Deeply Decoupled and Denoised Knowledge Distillation Trainer","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Haiduo Huang, Jiangcheng Song, Pengju Ren, Yadong Zhang","submitted_at":"2025-05-21T05:38:57Z","abstract_excerpt":"Recent advances in knowledge distillation have emphasized the importance of decoupling different knowledge components. While existing methods utilize momentum mechanisms to separate task-oriented and distillation gradients, they overlook the inherent conflict between target-class and non-target-class knowledge flows. Furthermore, low-confidence dark knowledge in non-target classes introduces noisy signals that hinder effective knowledge transfer. To address these limitations, we propose DeepKD, a novel training framework that integrates dual-level decoupling with adaptive denoising. First, thr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.15133","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/2505.15133/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":"2505.15133","created_at":"2026-07-05T11:06:41.106987+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.15133v1","created_at":"2026-07-05T11:06:41.106987+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.15133","created_at":"2026-07-05T11:06:41.106987+00:00"},{"alias_kind":"pith_short_12","alias_value":"4VCGMWYD4KBA","created_at":"2026-07-05T11:06:41.106987+00:00"},{"alias_kind":"pith_short_16","alias_value":"4VCGMWYD4KBARY5X","created_at":"2026-07-05T11:06:41.106987+00:00"},{"alias_kind":"pith_short_8","alias_value":"4VCGMWYD","created_at":"2026-07-05T11:06:41.106987+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/4VCGMWYD4KBARY5X7BCHWKQ52E","json":"https://pith.science/pith/4VCGMWYD4KBARY5X7BCHWKQ52E.json","graph_json":"https://pith.science/api/pith-number/4VCGMWYD4KBARY5X7BCHWKQ52E/graph.json","events_json":"https://pith.science/api/pith-number/4VCGMWYD4KBARY5X7BCHWKQ52E/events.json","paper":"https://pith.science/paper/4VCGMWYD"},"agent_actions":{"view_html":"https://pith.science/pith/4VCGMWYD4KBARY5X7BCHWKQ52E","download_json":"https://pith.science/pith/4VCGMWYD4KBARY5X7BCHWKQ52E.json","view_paper":"https://pith.science/paper/4VCGMWYD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.15133&json=true","fetch_graph":"https://pith.science/api/pith-number/4VCGMWYD4KBARY5X7BCHWKQ52E/graph.json","fetch_events":"https://pith.science/api/pith-number/4VCGMWYD4KBARY5X7BCHWKQ52E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4VCGMWYD4KBARY5X7BCHWKQ52E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4VCGMWYD4KBARY5X7BCHWKQ52E/action/storage_attestation","attest_author":"https://pith.science/pith/4VCGMWYD4KBARY5X7BCHWKQ52E/action/author_attestation","sign_citation":"https://pith.science/pith/4VCGMWYD4KBARY5X7BCHWKQ52E/action/citation_signature","submit_replication":"https://pith.science/pith/4VCGMWYD4KBARY5X7BCHWKQ52E/action/replication_record"}},"created_at":"2026-07-05T11:06:41.106987+00:00","updated_at":"2026-07-05T11:06:41.106987+00:00"}