{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:XFZRKPT7SEV2TFD5H7TYJA42TQ","short_pith_number":"pith:XFZRKPT7","schema_version":"1.0","canonical_sha256":"b973153e7f912ba9947d3fe784839a9c1440f2bf3c045a962675c201cb545fef","source":{"kind":"arxiv","id":"2605.20357","version":1},"attestation_state":"computed","paper":{"title":"Consistently Informative Soft-Label Temperature for Knowledge Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hoang-Chau Luong, Kaiqi Zhao, Lingwei Chen, Nghia Van Vo","submitted_at":"2026-05-19T18:11:54Z","abstract_excerpt":"Knowledge distillation (KD) transfers knowledge from a high-capacity teacher to a compact student by matching their predictive distributions, with temperature scaling serving as a central mechanism for smoothing teacher predictions and exposing informative \"dark knowledge\" beyond the hard label. However, the standard fixed-temperature design is inherently sample-agnostic. Since samples differ in logit scale and learning difficulty, a single global temperature produces teacher soft labels with highly inconsistent entropy: some predictions remain overly sharp and provide limited inter-class info"},"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":"2605.20357","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T18:11:54Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b495bcdc97d628e85838c36b924517a69eb3eb5c1594747e2d055453214ed623","abstract_canon_sha256":"60104747bd918e03288704a8109bf37aaa31d54b65ffeeea610bd29195498382"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:34.741629Z","signature_b64":"AN3tt8LM1yV6JT07/ecFeFhNkMFHmKmmckIzrsqaqFIsxIcfe5r63MkCSx+NClL4kRjn0Hqc9UtC/78VQ0yeDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b973153e7f912ba9947d3fe784839a9c1440f2bf3c045a962675c201cb545fef","last_reissued_at":"2026-05-21T01:04:34.741002Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:34.741002Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Consistently Informative Soft-Label Temperature for Knowledge Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hoang-Chau Luong, Kaiqi Zhao, Lingwei Chen, Nghia Van Vo","submitted_at":"2026-05-19T18:11:54Z","abstract_excerpt":"Knowledge distillation (KD) transfers knowledge from a high-capacity teacher to a compact student by matching their predictive distributions, with temperature scaling serving as a central mechanism for smoothing teacher predictions and exposing informative \"dark knowledge\" beyond the hard label. However, the standard fixed-temperature design is inherently sample-agnostic. Since samples differ in logit scale and learning difficulty, a single global temperature produces teacher soft labels with highly inconsistent entropy: some predictions remain overly sharp and provide limited inter-class info"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20357","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/2605.20357/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":"2605.20357","created_at":"2026-05-21T01:04:34.741094+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20357v1","created_at":"2026-05-21T01:04:34.741094+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20357","created_at":"2026-05-21T01:04:34.741094+00:00"},{"alias_kind":"pith_short_12","alias_value":"XFZRKPT7SEV2","created_at":"2026-05-21T01:04:34.741094+00:00"},{"alias_kind":"pith_short_16","alias_value":"XFZRKPT7SEV2TFD5","created_at":"2026-05-21T01:04:34.741094+00:00"},{"alias_kind":"pith_short_8","alias_value":"XFZRKPT7","created_at":"2026-05-21T01:04:34.741094+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/XFZRKPT7SEV2TFD5H7TYJA42TQ","json":"https://pith.science/pith/XFZRKPT7SEV2TFD5H7TYJA42TQ.json","graph_json":"https://pith.science/api/pith-number/XFZRKPT7SEV2TFD5H7TYJA42TQ/graph.json","events_json":"https://pith.science/api/pith-number/XFZRKPT7SEV2TFD5H7TYJA42TQ/events.json","paper":"https://pith.science/paper/XFZRKPT7"},"agent_actions":{"view_html":"https://pith.science/pith/XFZRKPT7SEV2TFD5H7TYJA42TQ","download_json":"https://pith.science/pith/XFZRKPT7SEV2TFD5H7TYJA42TQ.json","view_paper":"https://pith.science/paper/XFZRKPT7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20357&json=true","fetch_graph":"https://pith.science/api/pith-number/XFZRKPT7SEV2TFD5H7TYJA42TQ/graph.json","fetch_events":"https://pith.science/api/pith-number/XFZRKPT7SEV2TFD5H7TYJA42TQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XFZRKPT7SEV2TFD5H7TYJA42TQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XFZRKPT7SEV2TFD5H7TYJA42TQ/action/storage_attestation","attest_author":"https://pith.science/pith/XFZRKPT7SEV2TFD5H7TYJA42TQ/action/author_attestation","sign_citation":"https://pith.science/pith/XFZRKPT7SEV2TFD5H7TYJA42TQ/action/citation_signature","submit_replication":"https://pith.science/pith/XFZRKPT7SEV2TFD5H7TYJA42TQ/action/replication_record"}},"created_at":"2026-05-21T01:04:34.741094+00:00","updated_at":"2026-05-21T01:04:34.741094+00:00"}