{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:CX5GTD5SSKMODUPDXGBMPEHRAQ","short_pith_number":"pith:CX5GTD5S","schema_version":"1.0","canonical_sha256":"15fa698fb29298e1d1e3b982c790f1040ab12c05b24f9d4820ef6cab3b27400d","source":{"kind":"arxiv","id":"1909.01688","version":3},"attestation_state":"computed","paper":{"title":"Knowledge distillation for optimization of quantized deep neural networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Sungho Shin, Wonyong Sung, Yoonho Boo","submitted_at":"2019-09-04T10:47:03Z","abstract_excerpt":"Knowledge distillation (KD) is a very popular method for model size reduction. Recently, the technique is exploited for quantized deep neural networks (QDNNs) training as a way to restore the performance sacrificed by word-length reduction. KD, however, employs additional hyper-parameters, such as temperature, coefficient, and the size of teacher network for QDNN training. We analyze the effect of these hyper-parameters for QDNN optimization with KD. We find that these hyper-parameters are inter-related, and also introduce a simple and effective technique that reduces \\textit{coefficient} duri"},"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":"1909.01688","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-09-04T10:47:03Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"9a0752e6c8aa25094bbdf1f22792510d9508eadf16d64635526c3fd4986c5fb2","abstract_canon_sha256":"be3d0a836181b2c1a68dfe8be6bc2c88c6e2e08c900ca5eaff75a5c8f81707be"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:14:21.338376Z","signature_b64":"4YDZ7jND4fHCvuMDil/L1TYXyjNmQzrCHYyb44t5Q3cjy4HXXCoSlYBalxYn3qV0XkUw5GBSsPayMREN6cAGBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"15fa698fb29298e1d1e3b982c790f1040ab12c05b24f9d4820ef6cab3b27400d","last_reissued_at":"2026-07-05T00:14:21.338028Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:14:21.338028Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Knowledge distillation for optimization of quantized deep neural networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Sungho Shin, Wonyong Sung, Yoonho Boo","submitted_at":"2019-09-04T10:47:03Z","abstract_excerpt":"Knowledge distillation (KD) is a very popular method for model size reduction. Recently, the technique is exploited for quantized deep neural networks (QDNNs) training as a way to restore the performance sacrificed by word-length reduction. KD, however, employs additional hyper-parameters, such as temperature, coefficient, and the size of teacher network for QDNN training. We analyze the effect of these hyper-parameters for QDNN optimization with KD. We find that these hyper-parameters are inter-related, and also introduce a simple and effective technique that reduces \\textit{coefficient} duri"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1909.01688","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/1909.01688/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":"1909.01688","created_at":"2026-07-05T00:14:21.338087+00:00"},{"alias_kind":"arxiv_version","alias_value":"1909.01688v3","created_at":"2026-07-05T00:14:21.338087+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1909.01688","created_at":"2026-07-05T00:14:21.338087+00:00"},{"alias_kind":"pith_short_12","alias_value":"CX5GTD5SSKMO","created_at":"2026-07-05T00:14:21.338087+00:00"},{"alias_kind":"pith_short_16","alias_value":"CX5GTD5SSKMODUPD","created_at":"2026-07-05T00:14:21.338087+00:00"},{"alias_kind":"pith_short_8","alias_value":"CX5GTD5S","created_at":"2026-07-05T00:14:21.338087+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"1909.01688","citing_title":"Knowledge distillation for optimization of quantized deep neural networks","ref_index":2,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ","json":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ.json","graph_json":"https://pith.science/api/pith-number/CX5GTD5SSKMODUPDXGBMPEHRAQ/graph.json","events_json":"https://pith.science/api/pith-number/CX5GTD5SSKMODUPDXGBMPEHRAQ/events.json","paper":"https://pith.science/paper/CX5GTD5S"},"agent_actions":{"view_html":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ","download_json":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ.json","view_paper":"https://pith.science/paper/CX5GTD5S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1909.01688&json=true","fetch_graph":"https://pith.science/api/pith-number/CX5GTD5SSKMODUPDXGBMPEHRAQ/graph.json","fetch_events":"https://pith.science/api/pith-number/CX5GTD5SSKMODUPDXGBMPEHRAQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ/action/storage_attestation","attest_author":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ/action/author_attestation","sign_citation":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ/action/citation_signature","submit_replication":"https://pith.science/pith/CX5GTD5SSKMODUPDXGBMPEHRAQ/action/replication_record"}},"created_at":"2026-07-05T00:14:21.338087+00:00","updated_at":"2026-07-05T00:14:21.338087+00:00"}