Papers › Adversarial Examples Improve Image Recognition

Adversarial Examples Improve Image Recognition

21 Nov 2019CVPR 2020 6arXiv:1911.09665archive 2025-07-28

Cihang Xie, Mingxing Tan, Boqing Gong, Jiang Wang, Alan Yuille, Quoc V. Le

Adversarial examples are commonly viewed as a threat to ConvNets. Here we present an opposite perspective: adversarial examples can be used to improve image recognition models if harnessed in the right manner. We propose AdvProp, an enhanced adversarial training scheme which treats adversarial examples as additional examples, to prevent overfitting. Key to our method is the usage of a separate auxiliary batch norm for adversarial examples, as they have different underlying distributions to normal examples. We show that AdvProp improves a wide range of models on various image recognition tasks and performs better when the models are bigger. For instance, by applying AdvProp to the latest EfficientNet-B7 [28] on ImageNet, we achieve significant improvements on ImageNet (+0.7%), ImageNet-C (+6.5%), ImageNet-A (+7.0%), Stylized-ImageNet (+4.8%). With an enhanced EfficientNet-B8, our method achieves the state-of-the-art 85.5% ImageNet top-1 accuracy without extra data. This result even surpasses the best model in [20] which is trained with 3.5B Instagram images (~3000X more than ImageNet) and ~9.4X more parameters. Models are available at https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

No sample from the archive's official repository (tensorflow/tpu) is linked to this paper; the 2 samples below come from 1 community repository. Syntology harvested code from tensorflow/tpu for another paper.

tensorflow/tpu: Syntology's graph links these samples to Towards Understanding Why Label Smoothing Degrades Selective Classification and How to Fix It (arXiv:2403.14715; that paper's own run record: 9 ran (of which 2 constructed an object rather than computing a result; 6 with no instrument failure: 1 honoured, 0 violated, 5 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified); nothing checks that they implement this paper's method.

Syntology 2 code samples harvested from 1 repository linked to this paper: 2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified. Where Syntology's instrument failed: 2 where our contract draft was wrong; failures of Syntology's instrument, not of the code. Every run on this page is a failure of Syntology's instrument, not of the code.

By repository: community (archive-listed): 2 samples from 1 repository, 2 ran (0 with no instrument failure; 2 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 4 of its repositories that Syntology harvested for other papers:

No date is recorded for any of the 2 samples that ran: their levels were written before Syntology's graph kept the date.

tensorflow/tpu officialmentioned in papertfApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Towards Understanding Why Label Smoothing Degrades Selective Classification and How to Fix It (arXiv:2403.14715; that paper's own run record: 9 ran (of which 2 constructed an object rather than computing a result; 6 with no instrument failure: 1 honoured, 0 violated, 5 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified); nothing checks that they implement this paper's method. report
rwightman/pytorch-image-models mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights (arXiv:2006.08217; that paper's own run record: 4 ran (of which 1 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
yaoshiang/MobileNetV2-CIFAR-Cleverhans mentioned on GitHubSyntology: not harvested report
open-mmlab/mmclassification pytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Reversible Vision Transformers (arXiv:2302.04869; that paper's own run record: 24 ran (of which 5 constructed an object rather than computing a result; 17 with no instrument failure: 0 honoured, 0 violated, 17 with no contract checked; 7 where Syntology's instrument failed) · 5 unverified); nothing checks that they implement this paper's method. report
osmr/imgclsmob mxnetMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to ResNeSt: Split-Attention Networks (arXiv:2004.08955; that paper's own run record: 28 ran (of which 0 constructed an object rather than computing a result; 21 with no instrument failure: 0 honoured, 0 violated, 21 with no contract checked; 7 where Syntology's instrument failed) · 20 unverified); nothing checks that they implement this paper's method. report
tingxueronghua/pytorch-classification-advprop pytorchSyntology: 2 samples, 2 ran (2 instrument-failed) · 0 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

2 samples harvested: 2 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Every run on this page is a failure of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

2ran · our draft was wrong

Licence: 0 of the 2 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from tingxueronghua/pytorch-classification-advprop. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: 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. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

tingxueronghua/pytorch-classification-advprop community (archive-listed) 2 samples, 2 ran (2 instrument-failed) · 0 unverified
label_smoothing_cross_entropy imagenet.py ran · our draft was wrong MIT (permissive) · 6a259f7cc6738039 · report fixture: signature pred:Tensor/2/float, target:Tensor/1/int, classes:int/0/int, dim:int/0/int, reduction:str/0/str, smoothing:float/0/float · returned Tensor · level recorded, date not recorded
plot_overlap in 37 papers utils/logger.py ran · our draft was wrong MIT (permissive) · 50afa2863ffd9fde · report fixture: signature logger:MockLogger/0/?, names:NoneType/0/? · returned list · level recorded, date not recorded

Tasks

Domain GeneralizationImage Classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Domain Generalization VizWiz-Classification EfficientNet-B8 (advprop+autoaug) Accuracy - All Images 50.5 #4 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B8 (advprop+autoaug) Accuracy - Clean Images 53.2 #4 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B8 (advprop+autoaug) Accuracy - Corrupted Images 45.8 #4 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B7 (advprop+autoaug) Accuracy - All Images 49.7 #5 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B7 (advprop+autoaug) Accuracy - Clean Images 52 #5 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B7 (advprop+autoaug) Accuracy - Corrupted Images 45 #5 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B6 (advprop+autoaug) Accuracy - All Images 49.6 #6 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B6 (advprop+autoaug) Accuracy - Clean Images 53.2 #6 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B6 (advprop+autoaug) Accuracy - Corrupted Images 44.7 #6 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B5 (advprop+autoaug) Accuracy - All Images 49.1 #7 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B5 (advprop+autoaug) Accuracy - Clean Images 51.7 #7 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B5 (advprop+autoaug) Accuracy - Corrupted Images 44 #7 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B4 (advprop+autoaug) Accuracy - All Images 48.1 #10 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B4 (advprop+autoaug) Accuracy - Clean Images 51.4 #10 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B4 (advprop+autoaug) Accuracy - Corrupted Images 42.5 #10 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B3 (advprop+autoaug) Accuracy - All Images 45.5 #15 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B3 (advprop+autoaug) Accuracy - Clean Images 49.5 #15 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B3 (advprop+autoaug) Accuracy - Corrupted Images 39.8 #15 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B2 (advprop+autoaug) Accuracy - All Images 44.3 #19 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B2 (advprop+autoaug) Accuracy - Clean Images 48 #19 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B2 (advprop+autoaug) Accuracy - Corrupted Images 38.2 #19 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B1 (advprop+autoaug) Accuracy - All Images 42.4 #23 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B1 (advprop+autoaug) Accuracy - Clean Images 46.7 #23 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B1 (advprop+autoaug) Accuracy - Corrupted Images 36.2 #23 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B0 (advprop+autoaug) Accuracy - All Images 40.5 #34 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B0 (advprop+autoaug) Accuracy - Clean Images 44.9 #34 of 90 Archive leaderboard report
Domain Generalization VizWiz-Classification EfficientNet-B0 (advprop+autoaug) Accuracy - Corrupted Images 34.2 #34 of 90 Archive leaderboard report
Image Classification ImageNet AdvProp (EfficientNet-B8) Number of params 88M #224 of 1060 Archive leaderboard report
Image Classification ImageNet AdvProp (EfficientNet-B8) Top 1 Accuracy 85.5% #224 of 1060 Archive leaderboard report
Image Classification ImageNet AdvProp (EfficientNet-B7) Number of params 66M #249 of 1060 Archive leaderboard report
Image Classification ImageNet AdvProp (EfficientNet-B7) Top 1 Accuracy 85.2% #249 of 1060 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Introduced by this paper: AdvProp, Auxiliary Batch Normalization

1x1 ConvolutionAdvPropAutoAugmentAuxiliary Batch NormalizationAverage PoolingBatch NormalizationColorJitterConvolutionCutoutDense ConnectionsDepthwise ConvolutionDepthwise Separable ConvolutionDropoutEfficientNetInverted Residual BlockLSTMPointwise ConvolutionRMSPropReLUSigmoid ActivationSqueeze-and-Excitation BlockTanh ActivationWeight Decay

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections