{"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/faster-meta-update-strategy-for-noise-robust","title":"Faster Meta Update Strategy for Noise-Robust Deep Learning","arxiv_id":"2104.15092","date":"2021-04-30","proceeding":null,"authors":["Youjiang Xu","Linchao Zhu","Lu Jiang","Yi Yang"],"abstract":"It has been shown that deep neural networks are prone to overfitting on biased training data. Towards addressing this issue, meta-learning employs a meta model for correcting the training bias. Despite the promising performances, super slow training is currently the bottleneck in the meta learning approaches. In this paper, we introduce a novel Faster Meta Update Strategy (FaMUS) to replace the most expensive step in the meta gradient computation with a faster layer-wise approximation. We empirically find that FaMUS yields not only a reasonably accurate but also a low-variance approximation of the meta gradient. We conduct extensive experiments to verify the proposed method on two tasks. We show our method is able to save two-thirds of the training time while still maintaining the comparable or achieving even better generalization performance. In particular, our method achieves the state-of-the-art performance on both synthetic and realistic noisy labels, and obtains promising performance on long-tailed recognition on standard benchmarks.","url_abs":"https://arxiv.org/abs/2104.15092v1","url_pdf":"https://arxiv.org/pdf/2104.15092v1.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":"faster-meta-update-strategy-for-noise-robust","repo_url":"https://github.com/youjiangxu/FaMUS","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"deep-learning","task_name":"Deep Learning"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"learning-with-noisy-labels","task_name":"Learning with noisy labels"},{"task_slug":"meta-learning","task_name":"Meta-Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-classification-on-cifar-10-40-symmetric","task":"Image Classification","dataset":"CIFAR-10, 40% Symmetric Noise","model":"FaMUS","rank_in_archive_order":1,"of":2,"metrics":{"Percentage correct":"95.37"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-10-40-symmetric","task":"Image Classification","dataset":"CIFAR-10, 40% Symmetric Noise","model":"MentorMix","rank_in_archive_order":2,"of":2,"metrics":{"Percentage correct":"94.2"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-10-60-symmetric","task":"Image Classification","dataset":"CIFAR-10, 60% Symmetric Noise","model":"MentorMix","rank_in_archive_order":1,"of":2,"metrics":{"Percentage correct":"91.3"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-10-60-symmetric","task":"Image Classification","dataset":"CIFAR-10, 60% Symmetric Noise","model":"FaMUS","rank_in_archive_order":2,"of":2,"metrics":{"Percentage correct":"26.42"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-100-40","task":"Image Classification","dataset":"CIFAR-100, 40% Symmetric Noise","model":"FaMUS","rank_in_archive_order":1,"of":2,"metrics":{"Percentage correct":"75.91"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-100-40","task":"Image Classification","dataset":"CIFAR-100, 40% Symmetric Noise","model":"MentorMix","rank_in_archive_order":2,"of":2,"metrics":{"Percentage correct":"71.3"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-cifar-100-60","task":"Image Classification","dataset":"CIFAR-100, 60% Symmetric Noise","model":"MentorMix","rank_in_archive_order":1,"of":1,"metrics":{"Percentage correct":"64.6"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-red-miniimagenet-20","task":"Image Classification","dataset":"Red MiniImageNet 20% label noise","model":"FaMUS","rank_in_archive_order":5,"of":5,"metrics":{"Accuracy":"51.42"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-red-miniimagenet-40","task":"Image Classification","dataset":"Red MiniImageNet 40% label noise","model":"FaMUS","rank_in_archive_order":5,"of":5,"metrics":{"Accuracy":"48.06"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-red-miniimagenet-60","task":"Image Classification","dataset":"Red MiniImageNet 60% label noise","model":"FaMUS","rank_in_archive_order":4,"of":4,"metrics":{"Accuracy":"45.1"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-red-miniimagenet-80","task":"Image Classification","dataset":"Red MiniImageNet 80% label noise","model":"FaMUS","rank_in_archive_order":5,"of":5,"metrics":{"Accuracy":"35.5"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-mini-webvision-1-0","task":"Image Classification","dataset":"mini WebVision 1.0","model":"FaMUS","rank_in_archive_order":15,"of":47,"metrics":{"ImageNet Top-1 Accuracy":"77","ImageNet Top-5 Accuracy":"92.76","Top-1 Accuracy":"79.4","Top-5 Accuracy":"92.80"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2104.15092","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2104.15092"}},"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":"deterministic:regex_extraction","url":"https://github.com/youjiangxu/FaMUS","reach":null}],"summary":{"ran":1,"ran_draft_wrong":1,"unverified":5},"by_repo_kind":{"official":{"samples":6,"ran":1,"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":7,"samples":[{"code_sha256_prefix":"4a58e72ef633ee6c","entry":"MetaLinear","repo":"youjiangxu/FaMUS","repo_kind":"official","path":"FaMUS/grad_operator_layer.py","file_url":"https://github.com/youjiangxu/FaMUS/blob/HEAD/FaMUS/grad_operator_layer.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"4a58e72ef633ee6c"}},{"code_sha256_prefix":"a2f6a25906e664c5","entry":"to_var","repo":null,"repo_kind":null,"path":null,"file_url":null,"link_basis":"identical_code_first_harvested_elsewhere","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":null,"inline_ok":false,"mcp_get_code":{"code_sha256":"a2f6a25906e664c5"}},{"code_sha256_prefix":"c724d61684b4ae77","entry":"GradGumbelSoftmax","repo":"youjiangxu/FaMUS","repo_kind":"official","path":"FaMUS/grad_operator_layer.py","file_url":"https://github.com/youjiangxu/FaMUS/blob/HEAD/FaMUS/grad_operator_layer.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"c724d61684b4ae77"}},{"code_sha256_prefix":"5974b54bf3031173","entry":"MetaModule","repo":"youjiangxu/FaMUS","repo_kind":"official","path":"FaMUS/grad_operator_layer.py","file_url":"https://github.com/youjiangxu/FaMUS/blob/HEAD/FaMUS/grad_operator_layer.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"5974b54bf3031173"}},{"code_sha256_prefix":"f65d1dea35887b02","entry":"gumbel_softmax","repo":"youjiangxu/FaMUS","repo_kind":"official","path":"FaMUS/grad_operator_layer.py","file_url":"https://github.com/youjiangxu/FaMUS/blob/HEAD/FaMUS/grad_operator_layer.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"f65d1dea35887b02"}},{"code_sha256_prefix":"3685cb8a4375545e","entry":"gumbel_softmax_sample","repo":"youjiangxu/FaMUS","repo_kind":"official","path":"FaMUS/grad_operator_layer.py","file_url":"https://github.com/youjiangxu/FaMUS/blob/HEAD/FaMUS/grad_operator_layer.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3685cb8a4375545e"}},{"code_sha256_prefix":"4b67c548de39aee4","entry":"sample_gumbel","repo":"youjiangxu/FaMUS","repo_kind":"official","path":"FaMUS/grad_operator_layer.py","file_url":"https://github.com/youjiangxu/FaMUS/blob/HEAD/FaMUS/grad_operator_layer.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"4b67c548de39aee4"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}