Data to 7 Oct 2026

AutoAugment: Learning Augmentation Policies from Data

arXiv 1805.09501 ↗First posted 24 May 2018cs.CV cs.LG stat.MLcited by 728

Not on the map. Why?

Abstract (arXiv)

Data augmentation is an effective technique for improving the accuracy of modern image classifiers. However, current data augmentation implementations are manually designed. In this paper, we describe a simple procedure called AutoAugment to automatically search for improved data augmentation policies. In our implementation, we have designed a search space where a policy consists of many sub-policies, one of which is randomly chosen for each image in each mini-batch. A sub-policy consists of two operations, each operation being an image processing function such as translation, rotation, or shearing, and the probabilities and magnitudes with which the functions are applied. We use a search algorithm to find the best policy such that the neural network yields the highest validation accuracy on a target dataset. Our method achieves state-of-the-art accuracy on CIFAR-10, CIFAR-100, SVHN, and ImageNet (without additional data). On ImageNet, we attain a Top-1 accuracy of 83.5% which is 0.4% better than the previous record of 83.1%. On CIFAR-10, we achieve an error rate of 1.5%, which is 0.6% better than the previous state-of-the-art. Augmentation policies we find are transferable between datasets. The policy learned on ImageNet transfers well to achieve significant improvements on other datasets, such as Oxford Flowers, Caltech-101, Oxford-IIT Pets, FGVC Aircraft, and Stanford Cars.

Key facts. Cited by 728 papers in the map (data through 7 Oct 2026). First posted on arXiv 24 May 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 174; 2022: 169; 2023: 163; 2024: 134; 2025: 56; 2026: 32.

728 papers.

Sort:

By area:

2021MLP-Mixer: An all-MLP Architecture for Visioncited by 1,441
2021Membership Inference Attacks From First Principlescited by 839
2024Revisiting Feature Prediction for Learning Visual Representations from Videocited by 490
2022DeiT III: Revenge of the ViTcited by 417
2021A Survey of Data Augmentation Approaches for NLPcited by 408
2021Dynamic Neural Networks: A Surveycited by 405
2021Early Convolutions Help Transformers See Bettercited by 395
2021ResNet strikes back: An improved training procedure in timmcited by 365
2021Dataset Condensation with Differentiable Siamese Augmentationcited by 290
2021Improving Robustness using Generated Datacited by 233
2023Data-centric Artificial Intelligence: A Surveycited by 202
2021Fixing Data Augmentation to Improve Adversarial Robustnesscited by 201

38 more

2021Data Augmentation Can Improve Robustnesscited by 193
2022MERLOT Reserve: Neural Script Knowledge through Vision and Language and Soundcited by 184
2022Scaling Up Dataset Distillation to ImageNet-1K with Constant Memorycited by 180
2022Self-supervised Learning of Adversarial Example: Towards Good Generalizations for Deepfake Detectioncited by 172
2022Surrogate Gap Minimization Improves Sharpness-Aware Trainingcited by 171
2022A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunitiescited by 162
2021Revisiting ResNets: Improved Training and Scaling Strategiescited by 158
2023Reconstructing the Mind's Eye: fMRI-to-Image with Contrastive Learning and Diffusion Priorscited by 156
2021MoViNets: Mobile Video Networks for Efficient Video Recognitioncited by 135
2021TransFG: A Transformer Architecture for Fine-grained Recognitioncited by 131
2021PixMix: Dreamlike Pictures Comprehensively Improve Safety Measurescited by 131
2022Penalizing Gradient Norm for Efficiently Improving Generalization in Deep Learningcited by 122
2022Revealing the Dark Secrets of Masked Image Modelingcited by 117
2022Revisiting Weakly Supervised Pre-Training of Visual Perception Modelscited by 116
2021Adversarial Examples Make Strong Poisonscited by 115
2022MobileOne: An Improved One millisecond Mobile Backbonecited by 109
2021Learning Optimal Conformal Classifierscited by 108
2022An Empirical Study of Remote Sensing Pretrainingcited by 105
2021AutoFlow: Learning a Better Training Set for Optical Flowcited by 102
2021Contrastive Learning with Stronger Augmentationscited by 100
2021Neural Transformation Learning for Deep Anomaly Detection Beyond Imagescited by 98
2022SURF: Semi-supervised Reward Learning with Data Augmentation for Feedback-efficient Preference-based Reinforcement Learningcited by 96
2022Vision Transformers provably learn spatial structurecited by 92
2023Leaving Reality to Imagination: Robust Classification via Generated Datasetscited by 91
2021AugMax: Adversarial Composition of Random Augmentations for Robust Trainingcited by 89
2021Space-time Mixing Attention for Video Transformercited by 84
2023InceptionNeXt: When Inception Meets ConvNeXtcited by 84
2021ReSSL: Relational Self-Supervised Learning with Weak Augmentationcited by 83
2023DataDAM: Efficient Dataset Distillation with Attention Matchingcited by 83
2021A Simple Baseline for Semi-supervised Semantic Segmentation with Strong Data Augmentationcited by 79
2022Video K-Net: A Simple, Strong, and Unified Baseline for Video Segmentationcited by 79
2022Benchmarking Robustness of 3D Point Cloud Recognition Against Common Corruptionscited by 78
2024DiffuseMix: Label-Preserving Data Augmentation with Diffusion Modelscited by 78
2022UNICON: Combating Label Noise Through Uniform Selection and Contrastive Learningcited by 77
2021Zen-NAS: A Zero-Shot NAS for High-Performance Deep Image Recognitioncited by 71
2021On the effectiveness of adversarial training against common corruptionscited by 71
2022Teaching Structured Vision&Language Concepts to Vision&Language Modelscited by 71
2021Amplitude-Phase Recombination: Rethinking Robustness of Convolutional Neural Networks in Frequency Domaincited by 70

50 most cited of 728 All 728 (JSON)