{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:73ZTN7Q23WINDUTHOBY5GL56UI","short_pith_number":"pith:73ZTN7Q2","schema_version":"1.0","canonical_sha256":"fef336fe1add90d1d2677071d32fbea215215bb251eecdb576d4397eeeb08e7b","source":{"kind":"arxiv","id":"2105.03746","version":4},"attestation_state":"computed","paper":{"title":"Contrastive Attraction and Contrastive Repulsion for Representation Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chunyuan Li, Hao Zhang, Hongxia Yang, Huangjie Zheng, Ivor Tsang, Jiangchao Yao, Jingren Zhou, Mingyuan Zhou, Xu Chen, Ya Zhang","submitted_at":"2021-05-08T17:25:08Z","abstract_excerpt":"Contrastive learning (CL) methods effectively learn data representations in a self-supervision manner, where the encoder contrasts each positive sample over multiple negative samples via a one-vs-many softmax cross-entropy loss. By leveraging large amounts of unlabeled image data, recent CL methods have achieved promising results when pretrained on large-scale datasets, such as ImageNet. However, most of them consider the augmented views from the same instance are positive pairs, while views from other instances are negative ones. Such binary partition insufficiently considers the relation bet"},"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":"2105.03746","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-08T17:25:08Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"fff1e6dd7315050bc0841777f7e1c3672124491598d7fb3ffea5a64dd56e6526","abstract_canon_sha256":"79429845192f4a11b2df6401fb825fe5fe5c7e64844344743d5bce773c3a779a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:41:00.346691Z","signature_b64":"7rh25J39y9hLdTixHjl/zVa/cGCudB8BAtkklICyr3I3sZnjyNnwt/GNxHYoA8aQfkSSOdacZJa14FoTT+f0Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fef336fe1add90d1d2677071d32fbea215215bb251eecdb576d4397eeeb08e7b","last_reissued_at":"2026-07-05T06:41:00.346169Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:41:00.346169Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Contrastive Attraction and Contrastive Repulsion for Representation Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chunyuan Li, Hao Zhang, Hongxia Yang, Huangjie Zheng, Ivor Tsang, Jiangchao Yao, Jingren Zhou, Mingyuan Zhou, Xu Chen, Ya Zhang","submitted_at":"2021-05-08T17:25:08Z","abstract_excerpt":"Contrastive learning (CL) methods effectively learn data representations in a self-supervision manner, where the encoder contrasts each positive sample over multiple negative samples via a one-vs-many softmax cross-entropy loss. By leveraging large amounts of unlabeled image data, recent CL methods have achieved promising results when pretrained on large-scale datasets, such as ImageNet. However, most of them consider the augmented views from the same instance are positive pairs, while views from other instances are negative ones. Such binary partition insufficiently considers the relation bet"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.03746","kind":"arxiv","version":4},"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/2105.03746/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":"2105.03746","created_at":"2026-07-05T06:41:00.346239+00:00"},{"alias_kind":"arxiv_version","alias_value":"2105.03746v4","created_at":"2026-07-05T06:41:00.346239+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.03746","created_at":"2026-07-05T06:41:00.346239+00:00"},{"alias_kind":"pith_short_12","alias_value":"73ZTN7Q23WIN","created_at":"2026-07-05T06:41:00.346239+00:00"},{"alias_kind":"pith_short_16","alias_value":"73ZTN7Q23WINDUTH","created_at":"2026-07-05T06:41:00.346239+00:00"},{"alias_kind":"pith_short_8","alias_value":"73ZTN7Q2","created_at":"2026-07-05T06:41:00.346239+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.12451","citing_title":"FuTCR: Future-Targeted Contrast and Repulsion for Continual Panoptic Segmentation","ref_index":56,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12451","citing_title":"FuTCR: Future-Targeted Contrast and Repulsion for Continual Panoptic Segmentation","ref_index":56,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI","json":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI.json","graph_json":"https://pith.science/api/pith-number/73ZTN7Q23WINDUTHOBY5GL56UI/graph.json","events_json":"https://pith.science/api/pith-number/73ZTN7Q23WINDUTHOBY5GL56UI/events.json","paper":"https://pith.science/paper/73ZTN7Q2"},"agent_actions":{"view_html":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI","download_json":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI.json","view_paper":"https://pith.science/paper/73ZTN7Q2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2105.03746&json=true","fetch_graph":"https://pith.science/api/pith-number/73ZTN7Q23WINDUTHOBY5GL56UI/graph.json","fetch_events":"https://pith.science/api/pith-number/73ZTN7Q23WINDUTHOBY5GL56UI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI/action/storage_attestation","attest_author":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI/action/author_attestation","sign_citation":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI/action/citation_signature","submit_replication":"https://pith.science/pith/73ZTN7Q23WINDUTHOBY5GL56UI/action/replication_record"}},"created_at":"2026-07-05T06:41:00.346239+00:00","updated_at":"2026-07-05T06:41:00.346239+00:00"}