{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:6J5JI5EDSYFFKEWGW4JHMRP2R6","short_pith_number":"pith:6J5JI5ED","schema_version":"1.0","canonical_sha256":"f27a947483960a5512c6b7127645fa8f94ec5dd87aa3e23f1f5040a7ab12ca5a","source":{"kind":"arxiv","id":"2108.01513","version":3},"attestation_state":"computed","paper":{"title":"SphereFace2: Binary Classification is All You Need for Deep Face Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Adrian Weller, Bhiksha Raj, Rita Singh, Weiyang Liu, Yandong Wen","submitted_at":"2021-08-03T13:58:45Z","abstract_excerpt":"State-of-the-art deep face recognition methods are mostly trained with a softmax-based multi-class classification framework. Despite being popular and effective, these methods still have a few shortcomings that limit empirical performance. In this paper, we start by identifying the discrepancy between training and evaluation in the existing multi-class classification framework and then discuss the potential limitations caused by the \"competitive\" nature of softmax normalization. Motivated by these limitations, we propose a novel binary classification training framework, termed SphereFace2. In "},"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":"2108.01513","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-08-03T13:58:45Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"73a24cfb4e2f5766f7553b836cf9fea4c77a5cfeae0098691910c229d4a227e1","abstract_canon_sha256":"ecc22dd0dbd3b4d87f679339902d32bcb009a147d5ae81058f53596275a442c7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:12:50.200662Z","signature_b64":"Osw4MtgHYMrkwrMKfUBPcs7YSRTFetrpUk3WnQDF5dbCPhMW4/xBAyyz5hJiU/W1mbW5P7I/A8/yszgUx8q+DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f27a947483960a5512c6b7127645fa8f94ec5dd87aa3e23f1f5040a7ab12ca5a","last_reissued_at":"2026-07-05T04:12:50.200242Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:12:50.200242Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SphereFace2: Binary Classification is All You Need for Deep Face Recognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Adrian Weller, Bhiksha Raj, Rita Singh, Weiyang Liu, Yandong Wen","submitted_at":"2021-08-03T13:58:45Z","abstract_excerpt":"State-of-the-art deep face recognition methods are mostly trained with a softmax-based multi-class classification framework. Despite being popular and effective, these methods still have a few shortcomings that limit empirical performance. In this paper, we start by identifying the discrepancy between training and evaluation in the existing multi-class classification framework and then discuss the potential limitations caused by the \"competitive\" nature of softmax normalization. Motivated by these limitations, we propose a novel binary classification training framework, termed SphereFace2. In "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2108.01513","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/2108.01513/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":"2108.01513","created_at":"2026-07-05T04:12:50.200298+00:00"},{"alias_kind":"arxiv_version","alias_value":"2108.01513v3","created_at":"2026-07-05T04:12:50.200298+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2108.01513","created_at":"2026-07-05T04:12:50.200298+00:00"},{"alias_kind":"pith_short_12","alias_value":"6J5JI5EDSYFF","created_at":"2026-07-05T04:12:50.200298+00:00"},{"alias_kind":"pith_short_16","alias_value":"6J5JI5EDSYFFKEWG","created_at":"2026-07-05T04:12:50.200298+00:00"},{"alias_kind":"pith_short_8","alias_value":"6J5JI5ED","created_at":"2026-07-05T04:12:50.200298+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.13420","citing_title":"LVFace: Progressive Cluster Optimization for Large Vision Models in Face Recognition","ref_index":38,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6","json":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6.json","graph_json":"https://pith.science/api/pith-number/6J5JI5EDSYFFKEWGW4JHMRP2R6/graph.json","events_json":"https://pith.science/api/pith-number/6J5JI5EDSYFFKEWGW4JHMRP2R6/events.json","paper":"https://pith.science/paper/6J5JI5ED"},"agent_actions":{"view_html":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6","download_json":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6.json","view_paper":"https://pith.science/paper/6J5JI5ED","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2108.01513&json=true","fetch_graph":"https://pith.science/api/pith-number/6J5JI5EDSYFFKEWGW4JHMRP2R6/graph.json","fetch_events":"https://pith.science/api/pith-number/6J5JI5EDSYFFKEWGW4JHMRP2R6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6/action/storage_attestation","attest_author":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6/action/author_attestation","sign_citation":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6/action/citation_signature","submit_replication":"https://pith.science/pith/6J5JI5EDSYFFKEWGW4JHMRP2R6/action/replication_record"}},"created_at":"2026-07-05T04:12:50.200298+00:00","updated_at":"2026-07-05T04:12:50.200298+00:00"}