{"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/parametric-noise-injection-trainable","title":"Parametric Noise Injection: Trainable Randomness to Improve Deep Neural Network Robustness against Adversarial Attack","arxiv_id":"1811.09310","date":"2018-11-22","proceeding":"CVPR 2019 6","authors":["Adnan Siraj Rakin","Zhezhi He","Deliang Fan"],"abstract":"Recent development in the field of Deep Learning have exposed the underlying\nvulnerability of Deep Neural Network (DNN) against adversarial examples. In\nimage classification, an adversarial example is a carefully modified image that\nis visually imperceptible to the original image but can cause DNN model to\nmisclassify it. Training the network with Gaussian noise is an effective\ntechnique to perform model regularization, thus improving model robustness\nagainst input variation. Inspired by this classical method, we explore to\nutilize the regularization characteristic of noise injection to improve DNN's\nrobustness against adversarial attack. In this work, we propose\nParametric-Noise-Injection (PNI) which involves trainable Gaussian noise\ninjection at each layer on either activation or weights through solving the\nmin-max optimization problem, embedded with adversarial training. These\nparameters are trained explicitly to achieve improved robustness. To the best\nof our knowledge, this is the first work that uses trainable noise injection to\nimprove network robustness against adversarial attacks, rather than manually\nconfiguring the injected noise level through cross-validation. The extensive\nresults show that our proposed PNI technique effectively improves the\nrobustness against a variety of powerful white-box and black-box attacks such\nas PGD, C & W, FGSM, transferable attack and ZOO attack. Last but not the\nleast, PNI method improves both clean- and perturbed-data accuracy in\ncomparison to the state-of-the-art defense methods, which outperforms current\nunbroken PGD defense by 1.1 % and 6.8 % on clean test data and perturbed test\ndata respectively using Resnet-20 architecture.","url_abs":"http://arxiv.org/abs/1811.09310v1","url_pdf":"http://arxiv.org/pdf/1811.09310v1.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":"parametric-noise-injection-trainable","repo_url":"https://github.com/elliothe/CVPR_2019_PNI","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"adversarial-attack","task_name":"Adversarial Attack"},{"task_slug":"adversarial-defense","task_name":"Adversarial Defense"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1811.09310","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1811.09310"}},"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/elliothe/CVPR_2019_PNI","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"ran":4,"unverified":4},"by_repo_kind":{"official":{"samples":8,"ran":4,"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":"09b3d0b3ea772a30","entry":"convert_secs2time","repo":"elliothe/CVPR_2019_PNI","repo_kind":"official","path":"code/utils.py","file_url":"https://github.com/elliothe/CVPR_2019_PNI/blob/HEAD/code/utils.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"09b3d0b3ea772a30"}},{"code_sha256_prefix":"cd56898dd1531f85","entry":"vanilla_resnet20","repo":"elliothe/CVPR_2019_PNI","repo_kind":"official","path":"code/models/vanilla_models/vanilla_resnet_cifar.py","file_url":"https://github.com/elliothe/CVPR_2019_PNI/blob/HEAD/code/models/vanilla_models/vanilla_resnet_cifar.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"cd56898dd1531f85"}},{"code_sha256_prefix":"cab91481ef35184c","entry":"vanilla_resnet32","repo":"elliothe/CVPR_2019_PNI","repo_kind":"official","path":"code/models/vanilla_models/vanilla_resnet_cifar.py","file_url":"https://github.com/elliothe/CVPR_2019_PNI/blob/HEAD/code/models/vanilla_models/vanilla_resnet_cifar.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"cab91481ef35184c"}},{"code_sha256_prefix":"2acb1379b48508e4","entry":"vanilla_resnet44","repo":"elliothe/CVPR_2019_PNI","repo_kind":"official","path":"code/models/vanilla_models/vanilla_resnet_cifar.py","file_url":"https://github.com/elliothe/CVPR_2019_PNI/blob/HEAD/code/models/vanilla_models/vanilla_resnet_cifar.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"2acb1379b48508e4"}},{"code_sha256_prefix":"0fd256b4a9f1e67f","entry":"noise_resnet20","repo":"elliothe/CVPR_2019_PNI","repo_kind":"official","path":"code/models/noisy_resnet_cifar.py","file_url":"https://github.com/elliothe/CVPR_2019_PNI/blob/HEAD/code/models/noisy_resnet_cifar.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"0fd256b4a9f1e67f"}},{"code_sha256_prefix":"f8a863a7430e2faf","entry":"noise_resnet32","repo":"elliothe/CVPR_2019_PNI","repo_kind":"official","path":"code/models/noisy_resnet_cifar.py","file_url":"https://github.com/elliothe/CVPR_2019_PNI/blob/HEAD/code/models/noisy_resnet_cifar.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"f8a863a7430e2faf"}},{"code_sha256_prefix":"f9281016e42877da","entry":"noise_resnet44","repo":"elliothe/CVPR_2019_PNI","repo_kind":"official","path":"code/models/noisy_resnet_cifar.py","file_url":"https://github.com/elliothe/CVPR_2019_PNI/blob/HEAD/code/models/noisy_resnet_cifar.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"f9281016e42877da"}},{"code_sha256_prefix":"6de508671f95b11e","entry":"summary","repo":"elliothe/CVPR_2019_PNI","repo_kind":"official","path":"code/utils_/model_summary.py","file_url":"https://github.com/elliothe/CVPR_2019_PNI/blob/HEAD/code/utils_/model_summary.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"6de508671f95b11e"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}