{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:MPUEBW3PD2SNGNGBFEUAJ25OL2","short_pith_number":"pith:MPUEBW3P","schema_version":"1.0","canonical_sha256":"63e840db6f1ea4d334c1292804ebae5e9d760639fbcdce2fda09e34feb82b5eb","source":{"kind":"arxiv","id":"2108.05887","version":1},"attestation_state":"computed","paper":{"title":"Billion-Scale Pretraining with Vision Transformers for Multi-Task Visual Representations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Andrew Zhai, Dmitry Kislyuk, Dong Huk Park, Hao-Yu Wu, Josh Beal","submitted_at":"2021-08-12T17:58:56Z","abstract_excerpt":"Large-scale pretraining of visual representations has led to state-of-the-art performance on a range of benchmark computer vision tasks, yet the benefits of these techniques at extreme scale in complex production systems has been relatively unexplored. We consider the case of a popular visual discovery product, where these representations are trained with multi-task learning, from use-case specific visual understanding (e.g. skin tone classification) to general representation learning for all visual content (e.g. embeddings for retrieval). In this work, we describe how we (1) generate a datase"},"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":"2108.05887","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-08-12T17:58:56Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"9fd441e56c5405e345cf9f60931a7b60d2338d0a71c2934e85e8eb58b28c58d7","abstract_canon_sha256":"427b991e33fa555df491603e484c9eadcebb39215ff17d164825dc5bd8109835"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:05:23.540935Z","signature_b64":"7Ivt5lmGxiIp7cfv+NZ9+c6AGF7u5U5raIwAMnTtC+N20rio2ugRKNjZUtsYrTQqjgaPsZeIx4SHIrd4yZn6CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"63e840db6f1ea4d334c1292804ebae5e9d760639fbcdce2fda09e34feb82b5eb","last_reissued_at":"2026-07-05T03:05:23.540542Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:05:23.540542Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Billion-Scale Pretraining with Vision Transformers for Multi-Task Visual Representations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Andrew Zhai, Dmitry Kislyuk, Dong Huk Park, Hao-Yu Wu, Josh Beal","submitted_at":"2021-08-12T17:58:56Z","abstract_excerpt":"Large-scale pretraining of visual representations has led to state-of-the-art performance on a range of benchmark computer vision tasks, yet the benefits of these techniques at extreme scale in complex production systems has been relatively unexplored. We consider the case of a popular visual discovery product, where these representations are trained with multi-task learning, from use-case specific visual understanding (e.g. skin tone classification) to general representation learning for all visual content (e.g. embeddings for retrieval). In this work, we describe how we (1) generate a datase"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2108.05887","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/2108.05887/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":"2108.05887","created_at":"2026-07-05T03:05:23.540596+00:00"},{"alias_kind":"arxiv_version","alias_value":"2108.05887v1","created_at":"2026-07-05T03:05:23.540596+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2108.05887","created_at":"2026-07-05T03:05:23.540596+00:00"},{"alias_kind":"pith_short_12","alias_value":"MPUEBW3PD2SN","created_at":"2026-07-05T03:05:23.540596+00:00"},{"alias_kind":"pith_short_16","alias_value":"MPUEBW3PD2SNGNGB","created_at":"2026-07-05T03:05:23.540596+00:00"},{"alias_kind":"pith_short_8","alias_value":"MPUEBW3P","created_at":"2026-07-05T03:05:23.540596+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.21454","citing_title":"Visual Product Graph: Bridging Visual Products And Composite Images For End-to-End Style Recommendations","ref_index":2021,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2","json":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2.json","graph_json":"https://pith.science/api/pith-number/MPUEBW3PD2SNGNGBFEUAJ25OL2/graph.json","events_json":"https://pith.science/api/pith-number/MPUEBW3PD2SNGNGBFEUAJ25OL2/events.json","paper":"https://pith.science/paper/MPUEBW3P"},"agent_actions":{"view_html":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2","download_json":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2.json","view_paper":"https://pith.science/paper/MPUEBW3P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2108.05887&json=true","fetch_graph":"https://pith.science/api/pith-number/MPUEBW3PD2SNGNGBFEUAJ25OL2/graph.json","fetch_events":"https://pith.science/api/pith-number/MPUEBW3PD2SNGNGBFEUAJ25OL2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2/action/storage_attestation","attest_author":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2/action/author_attestation","sign_citation":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2/action/citation_signature","submit_replication":"https://pith.science/pith/MPUEBW3PD2SNGNGBFEUAJ25OL2/action/replication_record"}},"created_at":"2026-07-05T03:05:23.540596+00:00","updated_at":"2026-07-05T03:05:23.540596+00:00"}