Ctrl-A: Control-Driven Online Data Augmentation

TL;DR

Ctrl-A is an innovative online data augmentation method for image tasks that dynamically adjusts augmentation strengths during training using control theory principles, reducing manual tuning and improving performance.

Contribution

It introduces a control-theoretic approach for online adjustment of augmentation strengths, eliminating manual policy engineering and enhancing adaptability across image-vision tasks.

Findings

Achieves competitive results on CIFAR-10, CIFAR-100, and SVHN datasets.

Automatically suppresses detrimental augmentation styles during training.

Reduces need for manual tuning of augmentation policies.

Abstract

We introduce ControlAugment (Ctrl-A), an automated data augmentation algorithm for image-vision tasks, which incorporates principles from control theory for online adjustment of augmentation strength distributions during model training. Ctrl-A eliminates the need for initialization of individual augmentation strengths. Instead, augmentation strength distributions are dynamically, and individually, adapted during training based on a control-loop architecture and what we define as relative operation response curves. Using an operation-dependent update procedure provides Ctrl-A with the potential to suppress augmentation styles that negatively impact model performance, alleviating the need for manually engineering augmentation policies for new image-vision tasks. Experiments on the CIFAR-10, CIFAR-100, and SVHN-core benchmark datasets using the common WideResNet-28-10 architecture demonstrate that Ctrl-A is highly competitive with existing state-of-the-art data augmentation strategies.