{"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/dada-deep-adversarial-data-augmentation-for","title":"DADA: Deep Adversarial Data Augmentation for Extremely Low Data Regime Classification","arxiv_id":"1809.00981","date":"2018-08-29","proceeding":null,"authors":["Xiaofeng Zhang","Zhangyang Wang","Dong Liu","Qing Ling"],"abstract":"Deep learning has revolutionized the performance of classification, but\nmeanwhile demands sufficient labeled data for training. Given insufficient\ndata, while many techniques have been developed to help combat overfitting, the\nchallenge remains if one tries to train deep networks, especially in the\nill-posed extremely low data regimes: only a small set of labeled data are\navailable, and nothing -- including unlabeled data -- else. Such regimes arise\nfrom practical situations where not only data labeling but also data collection\nitself is expensive. We propose a deep adversarial data augmentation (DADA)\ntechnique to address the problem, in which we elaborately formulate data\naugmentation as a problem of training a class-conditional and supervised\ngenerative adversarial network (GAN). Specifically, a new discriminator loss is\nproposed to fit the goal of data augmentation, through which both real and\naugmented samples are enforced to contribute to and be consistent in finding\nthe decision boundaries. Tailored training techniques are developed\naccordingly. To quantitatively validate its effectiveness, we first perform\nextensive simulations to show that DADA substantially outperforms both\ntraditional data augmentation and a few GAN-based options. We then extend\nexperiments to three real-world small labeled datasets where existing data\naugmentation and/or transfer learning strategies are either less effective or\ninfeasible. All results endorse the superior capability of DADA in enhancing\nthe generalization ability of deep networks trained in practical extremely low\ndata regimes. Source code is available at\nhttps://github.com/SchafferZhang/DADA.","url_abs":"http://arxiv.org/abs/1809.00981v1","url_pdf":"http://arxiv.org/pdf/1809.00981v1.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":"dada-deep-adversarial-data-augmentation-for","repo_url":"https://github.com/SchafferZhang/DADA","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"dada-deep-adversarial-data-augmentation-for","repo_url":"https://github.com/daixiangzi/DADA-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"data-augmentation","task_name":"Data Augmentation"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":null,"task_name":"Generative Adversarial Network"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1809.00981","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1809.00981"}},"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/daixiangzi/DADA-pytorch","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/SchafferZhang/DADA","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":1,"samples":[{"code_sha256_prefix":"5a2aa745c05f5871","entry":"merge","repo":"SchafferZhang/DADA","repo_kind":"official","path":"cifar10-svhn/train_cifar_svhn.py","file_url":"https://github.com/SchafferZhang/DADA/blob/HEAD/cifar10-svhn/train_cifar_svhn.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":"5a2aa745c05f5871"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}