{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:VKM6NKHHDFRCQ4RKL4VN7CV4VZ","short_pith_number":"pith:VKM6NKHH","schema_version":"1.0","canonical_sha256":"aa99e6a8e7196228722a5f2adf8abcae7a418c36a55f77187f94a3a23e1c1656","source":{"kind":"arxiv","id":"2111.06377","version":3},"attestation_state":"computed","paper":{"title":"Masked Autoencoders Are Scalable Vision Learners","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Masked autoencoders learn scalable vision features by reconstructing heavily masked image patches.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Kaiming He, Piotr Doll\\'ar, Ross Girshick, Saining Xie, Xinlei Chen, Yanghao Li","submitted_at":"2021-11-11T18:46:40Z","abstract_excerpt":"This paper shows that masked autoencoders (MAE) are scalable self-supervised learners for computer vision. Our MAE approach is simple: we mask random patches of the input image and reconstruct the missing pixels. It is based on two core designs. First, we develop an asymmetric encoder-decoder architecture, with an encoder that operates only on the visible subset of patches (without mask tokens), along with a lightweight decoder that reconstructs the original image from the latent representation and mask tokens. Second, we find that masking a high proportion of the input image, e.g., 75%, yield"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2111.06377","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-11-11T18:46:40Z","cross_cats_sorted":[],"title_canon_sha256":"ea6c7158eab8cc332ab8f2aacdb3daf765d105415ff67a4b06ad26fe5613a9a1","abstract_canon_sha256":"6238adccafeedeef2db5d7eb1dd165dcc2a780a0709a21bf0621578993f81067"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:48.777379Z","signature_b64":"hCA0rE2tKv/dTdXV18kT9aVoOXLfSMGSOvI0A7K3DMjW8zp1kWH2hjdzkQokpxuieX+m7OSW5/DSH95PjOR0BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aa99e6a8e7196228722a5f2adf8abcae7a418c36a55f77187f94a3a23e1c1656","last_reissued_at":"2026-05-17T23:38:48.776710Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:48.776710Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Masked Autoencoders Are Scalable Vision Learners","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Masked autoencoders learn scalable vision features by reconstructing heavily masked image patches.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Kaiming He, Piotr Doll\\'ar, Ross Girshick, Saining Xie, Xinlei Chen, Yanghao Li","submitted_at":"2021-11-11T18:46:40Z","abstract_excerpt":"This paper shows that masked autoencoders (MAE) are scalable self-supervised learners for computer vision. Our MAE approach is simple: we mask random patches of the input image and reconstruct the missing pixels. It is based on two core designs. First, we develop an asymmetric encoder-decoder architecture, with an encoder that operates only on the visible subset of patches (without mask tokens), along with a lightweight decoder that reconstructs the original image from the latent representation and mask tokens. Second, we find that masking a high proportion of the input image, e.g., 75%, yield"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our scalable approach allows for learning high-capacity models that generalize well: e.g., a vanilla ViT-Huge model achieves the best accuracy (87.8%) among methods that use only ImageNet-1K data.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That masking a high proportion of the input (e.g. 75%) yields a nontrivial and meaningful self-supervisory task whose difficulty drives useful feature learning rather than trivial solutions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Masked autoencoders with asymmetric encoder-decoder and 75% masking ratio enable scalable self-supervised pre-training of vision transformers, achieving 87.8% ImageNet-1K accuracy with ViT-Huge using only unlabeled data.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Masked autoencoders learn scalable vision features by reconstructing heavily masked image patches.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"c817407f7174acfd1ba55553d8686b8d2df559320a3d10e33387caa6b0d6fec7"},"source":{"id":"2111.06377","kind":"arxiv","version":3},"verdict":{"id":"7a27ae39-4672-4f49-871f-747c0173bae4","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T06:49:33.669252Z","strongest_claim":"Our scalable approach allows for learning high-capacity models that generalize well: e.g., a vanilla ViT-Huge model achieves the best accuracy (87.8%) among methods that use only ImageNet-1K data.","one_line_summary":"Masked autoencoders with asymmetric encoder-decoder and 75% masking ratio enable scalable self-supervised pre-training of vision transformers, achieving 87.8% ImageNet-1K accuracy with ViT-Huge using only unlabeled data.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That masking a high proportion of the input (e.g. 75%) yields a nontrivial and meaningful self-supervisory task whose difficulty drives useful feature learning rather than trivial solutions.","pith_extraction_headline":"Masked autoencoders learn scalable vision features by reconstructing heavily masked image patches."},"references":{"count":74,"sample":[{"doi":"","year":2016,"title":"Layer Normalization","work_id":"20a2d720-0046-4c7c-bcd6-327ec8143f69","ref_index":1,"cited_arxiv_id":"1607.06450","is_internal_anchor":true},{"doi":"","year":2021,"title":"BEiT: BERT Pre-Training of Image Transformers","work_id":"d74eda3c-bf7e-45f1-a8f1-a0137ecca3f4","ref_index":2,"cited_arxiv_id":"2106.08254","is_internal_anchor":true},{"doi":"","year":1992,"title":"Self-organizing neural network that discovers surfaces in random-dot stereograms","work_id":"2a90a785-1b05-47c2-b9f2-70f36e69df81","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2020,"title":"Language mod- els are few-shot learners","work_id":"0b82323e-a2a3-43a2-8d5c-6b111825c722","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Emerging properties in self-supervised vision transformers","work_id":"b6e4b8f1-b87c-4a46-aa8f-ee63e646489a","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":74,"snapshot_sha256":"11d01219b550958a318eef7ab206f724197244f75f5cb537e2353204f99fc962","internal_anchors":5},"formal_canon":{"evidence_count":2,"snapshot_sha256":"f9f4c7850199158c1e43f43474bd607f847376ab3b62ff8e8839577725c6a4bd"},"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":"2111.06377","created_at":"2026-05-17T23:38:48.776812+00:00"},{"alias_kind":"arxiv_version","alias_value":"2111.06377v3","created_at":"2026-05-17T23:38:48.776812+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2111.06377","created_at":"2026-05-17T23:38:48.776812+00:00"},{"alias_kind":"pith_short_12","alias_value":"VKM6NKHHDFRC","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_16","alias_value":"VKM6NKHHDFRCQ4RK","created_at":"2026-05-18T12:33:33.725879+00:00"},{"alias_kind":"pith_short_8","alias_value":"VKM6NKHH","created_at":"2026-05-18T12:33:33.725879+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":31,"internal_anchor_count":31,"sample":[{"citing_arxiv_id":"2605.23045","citing_title":"The TIME Machine: On The Power of Motion for Efficient Perception","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2411.13549","citing_title":"KFC-W: Generating 3D-Consistent Videos from Unposed Internet Photos","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2507.01201","citing_title":"Escaping Plato's Cave: JAM for Aligning Independently Trained Vision and Language Models","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2510.16416","citing_title":"SSL4RL: Revisiting Self-supervised Learning as Intrinsic Reward for Visual-Language Reasoning","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2603.08155","citing_title":"C$^2$FG: Control Classifier-Free Guidance via Score Discrepancy Analysis","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2603.03190","citing_title":"Expectation and Acoustic Neural Network Representations Enhance Music Identification from Brain Activity","ref_index":68,"is_internal_anchor":true},{"citing_arxiv_id":"2605.15465","citing_title":"Toward World Modeling of Physiological Signals with Chaos-Theoretic Balancing and Latent Dynamics","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2510.18091","citing_title":"Accelerating Vision Transformers with Adaptive Patch Sizes","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2511.14427","citing_title":"Self-Supervised Multisensory Pretraining for Contact-Rich Robot Reinforcement Learning","ref_index":45,"is_internal_anchor":true},{"citing_arxiv_id":"2309.16671","citing_title":"Demystifying CLIP Data","ref_index":87,"is_internal_anchor":true},{"citing_arxiv_id":"2602.06382","citing_title":"Now You See That: Learning End-to-End Humanoid Locomotion from Raw Pixels","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2402.17762","citing_title":"Massive Activations in Large Language Models","ref_index":122,"is_internal_anchor":true},{"citing_arxiv_id":"2602.06912","citing_title":"PANC: Prior-Aware Normalized Cut via Anchor-Augmented Token Graphs","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2601.16175","citing_title":"Learning to Discover at Test Time","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2602.13780","citing_title":"Foundation Model-Driven Semantic Change Detection in Remote Sensing Imagery","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2603.08155","citing_title":"C$^2$FG: Control Classifier-Free Guidance via Score Discrepancy Analysis","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2205.01917","citing_title":"CoCa: Contrastive Captioners are Image-Text Foundation Models","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2211.14730","citing_title":"A Time Series is Worth 64 Words: Long-term Forecasting with Transformers","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2604.03522","citing_title":"Physics-Informed Transformer for Real-Time High-Fidelity Topology Optimization","ref_index":41,"is_internal_anchor":true},{"citing_arxiv_id":"2211.09085","citing_title":"Galactica: A Large Language Model for Science","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2404.08471","citing_title":"Revisiting Feature Prediction for Learning Visual Representations from Video","ref_index":248,"is_internal_anchor":true},{"citing_arxiv_id":"2604.19683","citing_title":"Mask World Model: Predicting What Matters for Robust Robot Policy Learning","ref_index":15,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00850","citing_title":"Earth System Foundation Model (ESFM): A unified framework for heterogeneous data integration and forecasting","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2604.10609","citing_title":"Self-supervised Pretraining of Cell Segmentation Models","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2604.10333","citing_title":"Zero-shot World Models Are Developmentally Efficient Learners","ref_index":44,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ","json":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ.json","graph_json":"https://pith.science/api/pith-number/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/graph.json","events_json":"https://pith.science/api/pith-number/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/events.json","paper":"https://pith.science/paper/VKM6NKHH"},"agent_actions":{"view_html":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ","download_json":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ.json","view_paper":"https://pith.science/paper/VKM6NKHH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2111.06377&json=true","fetch_graph":"https://pith.science/api/pith-number/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/graph.json","fetch_events":"https://pith.science/api/pith-number/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/action/storage_attestation","attest_author":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/action/author_attestation","sign_citation":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/action/citation_signature","submit_replication":"https://pith.science/pith/VKM6NKHHDFRCQ4RKL4VN7CV4VZ/action/replication_record"}},"created_at":"2026-05-17T23:38:48.776812+00:00","updated_at":"2026-05-17T23:38:48.776812+00:00"}