{"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"}