{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/adapting-imagenet-scale-models-to-complex","title":"If your data distribution shifts, use self-learning","arxiv_id":"2104.12928","date":"2021-04-27","proceeding":null,"authors":["Evgenia Rusak","Steffen Schneider","George Pachitariu","Luisa Eck","Peter Gehler","Oliver Bringmann","Wieland Brendel","Matthias Bethge"],"abstract":"We demonstrate that self-learning techniques like entropy minimization and pseudo-labeling are simple and effective at improving performance of a deployed computer vision model under systematic domain shifts. We conduct a wide range of large-scale experiments and show consistent improvements irrespective of the model architecture, the pre-training technique or the type of distribution shift. At the same time, self-learning is simple to use in practice because it does not require knowledge or access to the original training data or scheme, is robust to hyperparameter choices, is straight-forward to implement and requires only a few adaptation epochs. This makes self-learning techniques highly attractive for any practitioner who applies machine learning algorithms in the real world. We present state-of-the-art adaptation results on CIFAR10-C (8.5% error), ImageNet-C (22.0% mCE), ImageNet-R (17.4% error) and ImageNet-A (14.8% error), theoretically study the dynamics of self-supervised adaptation methods and propose a new classification dataset (ImageNet-D) which is challenging even with adaptation.","url_abs":"https://arxiv.org/abs/2104.12928v4","url_pdf":"https://arxiv.org/pdf/2104.12928v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"adapting-imagenet-scale-models-to-complex","repo_url":"https://github.com/bethgelab/robustness","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"robust-classification","task_name":"Robust classification"},{"task_slug":"self-learning","task_name":"Self-Learning"},{"task_slug":"unsupervised-domain-adaptation","task_name":"Unsupervised Domain Adaptation"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"depthwise-convolution","method_name":"Depthwise Convolution"},{"method_slug":"depthwise-separable-convolution","method_name":"Depthwise Separable Convolution"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"inverted-residual-block","method_name":"Inverted Residual Block"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"pointwise-convolution","method_name":"Pointwise Convolution"},{"method_slug":"rmsprop","method_name":"RMSProp"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"self-learning","method_name":"Self-Learning"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"squeeze-and-excitation-block","method_name":"Squeeze-and-Excitation Block"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-a","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-A","model":"EfficientNet-L2 NoisyStudent + RPL","rank_in_archive_order":1,"of":1,"metrics":{"Top 1 Error":"14.8"},"uses_additional_data":true},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"EfficientNet-L2+RPL","rank_in_archive_order":1,"of":16,"metrics":{"mean Corruption Error (mCE)":"22.0"},"uses_additional_data":true},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"EfficientNet-L2+ENT","rank_in_archive_order":2,"of":16,"metrics":{"mean Corruption Error (mCE)":"23.0"},"uses_additional_data":true},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"ResNeXt101 32x8d + DeepAug + Augmix + RPL","rank_in_archive_order":3,"of":16,"metrics":{"mean Corruption Error (mCE)":"34.8"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"ResNeXt101 32x8d + DeepAug + Augmix + ENT","rank_in_archive_order":4,"of":16,"metrics":{"mean Corruption Error (mCE)":"35.5"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"ResNeXt101 32x8d + IG-3.5B + ENT","rank_in_archive_order":7,"of":16,"metrics":{"mean Corruption Error (mCE)":"40.8"},"uses_additional_data":true},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"ResNeXt101 32x8d + IG-3.5B + RPL","rank_in_archive_order":8,"of":16,"metrics":{"mean Corruption Error (mCE)":"40.9"},"uses_additional_data":true},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"ResNeXt101 32x8d + RPL","rank_in_archive_order":9,"of":16,"metrics":{"mean Corruption Error (mCE)":"43.2"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"ResNeXt101 32x8d + ENT","rank_in_archive_order":10,"of":16,"metrics":{"mean Corruption Error (mCE)":"44.3"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"ResNet50 + RPL","rank_in_archive_order":13,"of":16,"metrics":{"mean Corruption Error (mCE)":"50.5"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-c","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-C","model":"ResNet50 + ENT","rank_in_archive_order":14,"of":16,"metrics":{"mean Corruption Error (mCE)":"51.6"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-r","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-R","model":"EfficientNet-L2 Noisy Student + RPL","rank_in_archive_order":2,"of":8,"metrics":{"Top 1 Error":"17.4"},"uses_additional_data":true},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-r","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-R","model":"EfficientNet-L2 Noisy Student + ENT","rank_in_archive_order":3,"of":8,"metrics":{"Top 1 Error":"19.7"},"uses_additional_data":true},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-r","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-R","model":"ResNet50 + RPL","rank_in_archive_order":6,"of":8,"metrics":{"Top 1 Error":"54.1"},"uses_additional_data":false},{"leaderboard":"/sota/unsupervised-domain-adaptation-on-imagenet-r","task":"Unsupervised Domain Adaptation","dataset":"ImageNet-R","model":"ResNet50 + ENT","rank_in_archive_order":7,"of":8,"metrics":{"Top 1 Error":"56.1"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2104.12928","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2104.12928"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/bethgelab/robustness","reach":null}],"summary":{"ran_honours":1,"ran_draft_wrong":1},"by_repo_kind":{"official":{"samples":2,"ran":2,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"57567db44fae299f","entry":"count_bn_layers","repo":"bethgelab/robustness","repo_kind":"official","path":"examples/batchnorm/bin/adapt_full.py","file_url":"https://github.com/bethgelab/robustness/blob/HEAD/examples/batchnorm/bin/adapt_full.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"57567db44fae299f"}},{"code_sha256_prefix":"b563e1eb405042ac","entry":"get_stages","repo":"bethgelab/robustness","repo_kind":"official","path":"examples/batchnorm/bin/adapt_full.py","file_url":"https://github.com/bethgelab/robustness/blob/HEAD/examples/batchnorm/bin/adapt_full.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"b563e1eb405042ac"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}