{"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/190407911","title":"REPAIR: Removing Representation Bias by Dataset Resampling","arxiv_id":"1904.07911","date":"2019-04-16","proceeding":"CVPR 2019 6","authors":["Yi Li","Nuno Vasconcelos"],"abstract":"Modern machine learning datasets can have biases for certain representations\nthat are leveraged by algorithms to achieve high performance without learning\nto solve the underlying task. This problem is referred to as \"representation\nbias\". The question of how to reduce the representation biases of a dataset is\ninvestigated and a new dataset REPresentAtion bIas Removal (REPAIR) procedure\nis proposed. This formulates bias minimization as an optimization problem,\nseeking a weight distribution that penalizes examples easy for a classifier\nbuilt on a given feature representation. Bias reduction is then equated to\nmaximizing the ratio between the classification loss on the reweighted dataset\nand the uncertainty of the ground-truth class labels. This is a minimax problem\nthat REPAIR solves by alternatingly updating classifier parameters and dataset\nresampling weights, using stochastic gradient descent. An experimental set-up\nis also introduced to measure the bias of any dataset for a given\nrepresentation, and the impact of this bias on the performance of recognition\nmodels. Experiments with synthetic and action recognition data show that\ndataset REPAIR can significantly reduce representation bias, and lead to\nimproved generalization of models trained on REPAIRed datasets. The tools used\nfor characterizing representation bias, and the proposed dataset REPAIR\nalgorithm, are available at https://github.com/JerryYLi/Dataset-REPAIR/.","url_abs":"http://arxiv.org/abs/1904.07911v1","url_pdf":"http://arxiv.org/pdf/1904.07911v1.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":"190407911","repo_url":"https://github.com/JerryYLi/Dataset-REPAIR","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"}],"methods":[{"method_slug":"repair","method_name":"Repair"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1904.07911","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1904.07911"}},"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/JerryYLi/Dataset-REPAIR","reach":null}],"summary":{"unverified":1},"by_repo_kind":{"official":{"samples":1,"ran":0,"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":"d85e7ab58a6b3906","entry":"repair","repo":"JerryYLi/Dataset-REPAIR","repo_kind":"official","path":"utils/repair.py","file_url":"https://github.com/JerryYLi/Dataset-REPAIR/blob/HEAD/utils/repair.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d85e7ab58a6b3906"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}