{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:JAMNFTZQRXOFBNKV42CYU6NYHY","short_pith_number":"pith:JAMNFTZQ","schema_version":"1.0","canonical_sha256":"4818d2cf308ddc50b555e6858a79b83e03fb0a65b2f675eb09728f56798e685f","source":{"kind":"arxiv","id":"2206.04667","version":2},"attestation_state":"computed","paper":{"title":"Extreme Masking for Learning Instance and Distributed Visual Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Stephen Lin, Xiao Sun, Zhirong Wu, Zihang Lai","submitted_at":"2022-06-09T17:59:43Z","abstract_excerpt":"The paper presents a scalable approach for learning spatially distributed visual representations over individual tokens and a holistic instance representation simultaneously. We use self-attention blocks to represent spatially distributed tokens, followed by cross-attention blocks to aggregate the holistic image instance. The core of the approach is the use of extremely large token masking (75\\%-90\\%) as the data augmentation for supervision. Our model, named ExtreMA, follows the plain BYOL approach where the instance representation from the unmasked subset is trained to predict that from the "},"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":"2206.04667","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2022-06-09T17:59:43Z","cross_cats_sorted":[],"title_canon_sha256":"f33b8b8afecf228b8b21c1dc7a4a224061e3597d24ef55d44f5d516038243f66","abstract_canon_sha256":"9f3ace6616166fd57d90a25b9d46dee0d889f900bfa71a796079299b3428b30e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:49:13.645279Z","signature_b64":"tMhlqf2vZ4VWiXKEokZpeRgNvS2R24JQlwPigzzoREKl/1IILu1fvO3Pp5jra+xpFNfGhydz/pbJedrAb+CnBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4818d2cf308ddc50b555e6858a79b83e03fb0a65b2f675eb09728f56798e685f","last_reissued_at":"2026-07-05T05:49:13.644706Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:49:13.644706Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Extreme Masking for Learning Instance and Distributed Visual Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Stephen Lin, Xiao Sun, Zhirong Wu, Zihang Lai","submitted_at":"2022-06-09T17:59:43Z","abstract_excerpt":"The paper presents a scalable approach for learning spatially distributed visual representations over individual tokens and a holistic instance representation simultaneously. We use self-attention blocks to represent spatially distributed tokens, followed by cross-attention blocks to aggregate the holistic image instance. The core of the approach is the use of extremely large token masking (75\\%-90\\%) as the data augmentation for supervision. Our model, named ExtreMA, follows the plain BYOL approach where the instance representation from the unmasked subset is trained to predict that from the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2206.04667","kind":"arxiv","version":2},"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/2206.04667/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":"2206.04667","created_at":"2026-07-05T05:49:13.644789+00:00"},{"alias_kind":"arxiv_version","alias_value":"2206.04667v2","created_at":"2026-07-05T05:49:13.644789+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2206.04667","created_at":"2026-07-05T05:49:13.644789+00:00"},{"alias_kind":"pith_short_12","alias_value":"JAMNFTZQRXOF","created_at":"2026-07-05T05:49:13.644789+00:00"},{"alias_kind":"pith_short_16","alias_value":"JAMNFTZQRXOFBNKV","created_at":"2026-07-05T05:49:13.644789+00:00"},{"alias_kind":"pith_short_8","alias_value":"JAMNFTZQ","created_at":"2026-07-05T05:49:13.644789+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/JAMNFTZQRXOFBNKV42CYU6NYHY","json":"https://pith.science/pith/JAMNFTZQRXOFBNKV42CYU6NYHY.json","graph_json":"https://pith.science/api/pith-number/JAMNFTZQRXOFBNKV42CYU6NYHY/graph.json","events_json":"https://pith.science/api/pith-number/JAMNFTZQRXOFBNKV42CYU6NYHY/events.json","paper":"https://pith.science/paper/JAMNFTZQ"},"agent_actions":{"view_html":"https://pith.science/pith/JAMNFTZQRXOFBNKV42CYU6NYHY","download_json":"https://pith.science/pith/JAMNFTZQRXOFBNKV42CYU6NYHY.json","view_paper":"https://pith.science/paper/JAMNFTZQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2206.04667&json=true","fetch_graph":"https://pith.science/api/pith-number/JAMNFTZQRXOFBNKV42CYU6NYHY/graph.json","fetch_events":"https://pith.science/api/pith-number/JAMNFTZQRXOFBNKV42CYU6NYHY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JAMNFTZQRXOFBNKV42CYU6NYHY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JAMNFTZQRXOFBNKV42CYU6NYHY/action/storage_attestation","attest_author":"https://pith.science/pith/JAMNFTZQRXOFBNKV42CYU6NYHY/action/author_attestation","sign_citation":"https://pith.science/pith/JAMNFTZQRXOFBNKV42CYU6NYHY/action/citation_signature","submit_replication":"https://pith.science/pith/JAMNFTZQRXOFBNKV42CYU6NYHY/action/replication_record"}},"created_at":"2026-07-05T05:49:13.644789+00:00","updated_at":"2026-07-05T05:49:13.644789+00:00"}