{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:B4ZIHHFQJ6CARBNQX6ZAPDIHLT","short_pith_number":"pith:B4ZIHHFQ","schema_version":"1.0","canonical_sha256":"0f32839cb04f840885b0bfb2078d075ce78e8dc5c0bced071a0fe4ef5520f0a3","source":{"kind":"arxiv","id":"2302.10899","version":3},"attestation_state":"computed","paper":{"title":"Feature Affinity Assisted Knowledge Distillation and Quantization of Deep Neural Networks on Label-Free Data","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.IT","cs.NA","math.IT","math.NA"],"primary_cat":"cs.LG","authors_text":"Biao Yang, Jack Xin, Penghang Yin, Yingyong Qi, Zhijian Li","submitted_at":"2023-02-10T01:00:49Z","abstract_excerpt":"In this paper, we propose a feature affinity (FA) assisted knowledge distillation (KD) method to improve quantization-aware training of deep neural networks (DNN). The FA loss on intermediate feature maps of DNNs plays the role of teaching middle steps of a solution to a student instead of only giving final answers in the conventional KD where the loss acts on the network logits at the output level. Combining logit loss and FA loss, we found that the quantized student network receives stronger supervision than from the labeled ground-truth data. The resulting FAQD is capable of compressing mod"},"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":"2302.10899","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2023-02-10T01:00:49Z","cross_cats_sorted":["cs.AI","cs.IT","cs.NA","math.IT","math.NA"],"title_canon_sha256":"aef6a5df4b296a6cfccc48a1cc15b136c41e58d14763b15631b8e7c7b3c095f7","abstract_canon_sha256":"c1b1a6507c469373ce92e1513c36416f360a8565d04bcc4c3aedcccec129268c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:42:42.508871Z","signature_b64":"zRTUZfEz7rL1+yTJBTG4Jxiu2zr+MdaF6Soxx8lwMxSl4a3XOso8QTci7DWSLsE1VF1G95GydoDlV54xdQxNDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0f32839cb04f840885b0bfb2078d075ce78e8dc5c0bced071a0fe4ef5520f0a3","last_reissued_at":"2026-07-05T06:42:42.508411Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:42:42.508411Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Feature Affinity Assisted Knowledge Distillation and Quantization of Deep Neural Networks on Label-Free Data","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.IT","cs.NA","math.IT","math.NA"],"primary_cat":"cs.LG","authors_text":"Biao Yang, Jack Xin, Penghang Yin, Yingyong Qi, Zhijian Li","submitted_at":"2023-02-10T01:00:49Z","abstract_excerpt":"In this paper, we propose a feature affinity (FA) assisted knowledge distillation (KD) method to improve quantization-aware training of deep neural networks (DNN). The FA loss on intermediate feature maps of DNNs plays the role of teaching middle steps of a solution to a student instead of only giving final answers in the conventional KD where the loss acts on the network logits at the output level. Combining logit loss and FA loss, we found that the quantized student network receives stronger supervision than from the labeled ground-truth data. The resulting FAQD is capable of compressing mod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.10899","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/2302.10899/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":"2302.10899","created_at":"2026-07-05T06:42:42.508482+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.10899v3","created_at":"2026-07-05T06:42:42.508482+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.10899","created_at":"2026-07-05T06:42:42.508482+00:00"},{"alias_kind":"pith_short_12","alias_value":"B4ZIHHFQJ6CA","created_at":"2026-07-05T06:42:42.508482+00:00"},{"alias_kind":"pith_short_16","alias_value":"B4ZIHHFQJ6CARBNQ","created_at":"2026-07-05T06:42:42.508482+00:00"},{"alias_kind":"pith_short_8","alias_value":"B4ZIHHFQ","created_at":"2026-07-05T06:42:42.508482+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/B4ZIHHFQJ6CARBNQX6ZAPDIHLT","json":"https://pith.science/pith/B4ZIHHFQJ6CARBNQX6ZAPDIHLT.json","graph_json":"https://pith.science/api/pith-number/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/graph.json","events_json":"https://pith.science/api/pith-number/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/events.json","paper":"https://pith.science/paper/B4ZIHHFQ"},"agent_actions":{"view_html":"https://pith.science/pith/B4ZIHHFQJ6CARBNQX6ZAPDIHLT","download_json":"https://pith.science/pith/B4ZIHHFQJ6CARBNQX6ZAPDIHLT.json","view_paper":"https://pith.science/paper/B4ZIHHFQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.10899&json=true","fetch_graph":"https://pith.science/api/pith-number/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/graph.json","fetch_events":"https://pith.science/api/pith-number/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/action/storage_attestation","attest_author":"https://pith.science/pith/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/action/author_attestation","sign_citation":"https://pith.science/pith/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/action/citation_signature","submit_replication":"https://pith.science/pith/B4ZIHHFQJ6CARBNQX6ZAPDIHLT/action/replication_record"}},"created_at":"2026-07-05T06:42:42.508482+00:00","updated_at":"2026-07-05T06:42:42.508482+00:00"}