Smart Augmentation Learning an Optimal Data Augmentation Strategy

By Joseph Lemley, Shabab Bazrafkan, Peter M. Corcoran. 2017

403 citations recorded by the source corpus.

A recurring problem faced when training neural networks is that there is typically not enough data to maximize the generalization capability of deep neural networks. There are many techniques to address this, including data augmentation, dropout, and transfer learning. In this paper, we introduce an additional method, which we call smart augmentation and we show how to use it to increase the accuracy and reduce over fitting on a target network. Smart augmentation works, by creating a network that learns how to generate augmented data during the training process of a target network in a way that reduces that networks loss. This allows us to learn augmentations that minimize the error of that network. Smart augmentation has shown the potential to increase accuracy by demonstrably significant measures on all data sets tested. In addition, it has shown potential to achieve similar or improved performance levels with significantly smaller network sizes in a number of tested cases.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

Citations in the corpus

Stochastic Optimization of Plain Convolutional Neural Networks with Simple Methods (2020, background): “Lemley et al.came up with a learned end-to-end approach called Smart Augmentation [6] instead of relying on hard-coded transformations.”
Randaugment: Practical automated data augmentation with a reduced search space (2019, background): “For example, Smart Augmentation learns a network that merges two or more samples from the same class to generate new data [18].”
AutoAugment: Learning Augmentation Policies from Data (2018, background): “Previous attempts at learned data augmentations include Smart Augmentation, which proposed a network that automatically generates augmented data by merging two or more samples from the same class [41].”
Adversarial AutoAugment (2019, methodology): “Hence, several works (Lemley et al., 2017; Cubuk et al., 2018; Lin et al., 2019; Ho et al., 2019) have attempted to automatically learn data augmentation policies.”
Automatically Learning Data Augmentation Policies for Dialogue Tasks (2019, background): “There has been extensive work that employs data augmentation in both computer vision (Simard et al., 2003; Krizhevsky et al., 2012; Cireşan et al., 2012; Wan et al., 2013; Sato et al., 2015; DeVries and Taylor, 2017; Tran et al., 2017; Lemley et al., 2017) and NLP (Fürstenau and Lapata, 2009; Sennri”
Population Based Augmentation: Efficient Learning of Augmentation Policy Schedules (2019, neutral): “Additionally, (Lemley et al., 2017) uses a network to combine pairs of images to train a target network, and (DeVries & Taylor, 2017) injects noise and interpolates images in an autoencoder learned feature space.”
Fast AutoAugment (2019, background): “Search of Augmentation Policies per Class Taking advantage of the fact that the algorithm is efficient, we experimented with searching for augmentation policies per class in CIFAR-100 with Wide-ResNet-40-2 Model.”
Albumentations: fast and flexible image augmentations (2018, methodology): “binations of ﬂipping, rotating, scaling, and cropping. Different domains, imaging modalities, and tasks may beneﬁt from a wide range of different amounts and combinations of image transformations [4]–[6]. For example, very extensive image augmentations are typical in medical image analysis, where da”
A survey on Image Data Augmentation for Deep Learning (2019, methodology): “Network-A in Smart Augmentation given the right images as input [37]”
Adversarial Examples Improve Image Recognition (2019, background): “Recent works also demonstrate that it is possible to learn data augmentation policies automatically for achieving better performance on image classification [3, 4, 19, 21, 52] and object detection [53, 4].”
Learning Data Augmentation Strategies for Object Detection (2019, background): “For example, Smart Augmentation uses a network that generates new data by merging two or more samples from the same class [22].”
Research on data augmentation for image classification based on convolution neural networks (2017, methodology): “Data augmentation, which refers to the process of creating new similar samples to the training set, can be regarded as one kind of regularization technology[6].”
Data augmentation instead of explicit regularization (2018, background): “Very interestingly, a promising avenue for future research has been set by recently proposed models that automatically learn the data transformations (Hauberg et al., 2016; Lemley et al., 2017; Ratner et al., 2017; Antoniou et al., 2017).”
Domain Randomization and Pyramid Consistency: Simulation-to-Real Generalization Without Accessing Target Domain Data (2019, background): “In [25], a network is proposed to automatically generate augmented data by merging two or more samples from the same class.”
A comprehensive survey of recent trends in deep learning for digital images augmentation (2021, methodology): “By utilizing another face dataset, the findings changed to 95.66% (Lemley et al. 2017).”
MixUp as Locally Linear Out-Of-Manifold Regularization (2018, background): “(Lemley, Bazrafkan, and Corcoran 2017; Cubuk et al. 2018).”
Data Augmentation Revisited: Rethinking the Distribution Gap between Clean and Augmented Data (2019, background): “Learning more sensible data augmentations for speciﬁc datasets has been explored [27, 32, 39].”
Enhancing Performance of Deep Learning Models with different Data Augmentation Techniques: A Survey (2020, methodology): “In [14] the authors have presented new method called smart augmentation to increase the precision and to solve the problem of overfitting of the target network.”