{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:G5HQKL7I4LVOM2LMSFXF26EIX3","short_pith_number":"pith:G5HQKL7I","schema_version":"1.0","canonical_sha256":"374f052fe8e2eae6696c916e5d7888beef4b65f6bb6c5d6727951a3a955b4426","source":{"kind":"arxiv","id":"2202.03670","version":2},"attestation_state":"computed","paper":{"title":"How to Understand Masked Autoencoders","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"David A. Clifton, Peng Xu, Shuhao Cao","submitted_at":"2022-02-08T06:15:07Z","abstract_excerpt":"\"Masked Autoencoders (MAE) Are Scalable Vision Learners\" revolutionizes the self-supervised learning method in that it not only achieves the state-of-the-art for image pre-training, but is also a milestone that bridges the gap between visual and linguistic masked autoencoding (BERT-style) pre-trainings. However, to our knowledge, to date there are no theoretical perspectives to explain the powerful expressivity of MAE. In this paper, we, for the first time, propose a unified theoretical framework that provides a mathematical understanding for MAE. Specifically, we explain the patch-based atten"},"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":"2202.03670","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2022-02-08T06:15:07Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"5308dcc1670558f34057331e6fee1c779d015a40ae1d196bfe6cfa6548d68b88","abstract_canon_sha256":"bfe211a93e33bd290cbe44c6872998517bf00890698f306117cce6b63b0e2e3b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:55:32.926771Z","signature_b64":"9Pyquu1LqC+xWqxebn9k2de++847D4Xohxw3gHpW9NylXVmviUEAaAxO0Hb2qLAAu/DStjg30XaTrOUz9ttEBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"374f052fe8e2eae6696c916e5d7888beef4b65f6bb6c5d6727951a3a955b4426","last_reissued_at":"2026-07-05T03:55:32.926252Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:55:32.926252Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How to Understand Masked Autoencoders","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"David A. Clifton, Peng Xu, Shuhao Cao","submitted_at":"2022-02-08T06:15:07Z","abstract_excerpt":"\"Masked Autoencoders (MAE) Are Scalable Vision Learners\" revolutionizes the self-supervised learning method in that it not only achieves the state-of-the-art for image pre-training, but is also a milestone that bridges the gap between visual and linguistic masked autoencoding (BERT-style) pre-trainings. However, to our knowledge, to date there are no theoretical perspectives to explain the powerful expressivity of MAE. In this paper, we, for the first time, propose a unified theoretical framework that provides a mathematical understanding for MAE. Specifically, we explain the patch-based atten"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.03670","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/2202.03670/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":"2202.03670","created_at":"2026-07-05T03:55:32.926326+00:00"},{"alias_kind":"arxiv_version","alias_value":"2202.03670v2","created_at":"2026-07-05T03:55:32.926326+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.03670","created_at":"2026-07-05T03:55:32.926326+00:00"},{"alias_kind":"pith_short_12","alias_value":"G5HQKL7I4LVO","created_at":"2026-07-05T03:55:32.926326+00:00"},{"alias_kind":"pith_short_16","alias_value":"G5HQKL7I4LVOM2LM","created_at":"2026-07-05T03:55:32.926326+00:00"},{"alias_kind":"pith_short_8","alias_value":"G5HQKL7I","created_at":"2026-07-05T03:55:32.926326+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":177,"is_internal_anchor":false},{"citing_arxiv_id":"2607.02203","citing_title":"Self-explainable Operator Learning for Discovering Spatial Patterns in Functional Data","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22873","citing_title":"SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning","ref_index":176,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3","json":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3.json","graph_json":"https://pith.science/api/pith-number/G5HQKL7I4LVOM2LMSFXF26EIX3/graph.json","events_json":"https://pith.science/api/pith-number/G5HQKL7I4LVOM2LMSFXF26EIX3/events.json","paper":"https://pith.science/paper/G5HQKL7I"},"agent_actions":{"view_html":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3","download_json":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3.json","view_paper":"https://pith.science/paper/G5HQKL7I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2202.03670&json=true","fetch_graph":"https://pith.science/api/pith-number/G5HQKL7I4LVOM2LMSFXF26EIX3/graph.json","fetch_events":"https://pith.science/api/pith-number/G5HQKL7I4LVOM2LMSFXF26EIX3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3/action/storage_attestation","attest_author":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3/action/author_attestation","sign_citation":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3/action/citation_signature","submit_replication":"https://pith.science/pith/G5HQKL7I4LVOM2LMSFXF26EIX3/action/replication_record"}},"created_at":"2026-07-05T03:55:32.926326+00:00","updated_at":"2026-07-05T03:55:32.926326+00:00"}