Invariant backpropagation: how to train a transformation-invariant neural network
This work addresses the need for transformation-invariant classifiers in machine learning, particularly for image classification tasks, but it is incremental as it builds on and improves existing robustness methods.
The authors tackled the problem of training neural networks to be robust to small input variations, proposing two backpropagation extensions that enforce robustness, with the second method showing better results but being slightly slower. They achieved improved classification accuracy and robustness on MNIST and CIFAR-10 datasets, with concrete comparisons to existing methods like Tangent BP and Adversarial Training.
In many classification problems a classifier should be robust to small variations in the input vector. This is a desired property not only for particular transformations, such as translation and rotation in image classification problems, but also for all others for which the change is small enough to retain the object perceptually indistinguishable. We propose two extensions of the backpropagation algorithm that train a neural network to be robust to variations in the feature vector. While the first of them enforces robustness of the loss function to all variations, the second method trains the predictions to be robust to a particular variation which changes the loss function the most. The second methods demonstrates better results, but is slightly slower. We analytically compare the proposed algorithm with two the most similar approaches (Tangent BP and Adversarial Training), and propose their fast versions. In the experimental part we perform comparison of all algorithms in terms of classification accuracy and robustness to noise on MNIST and CIFAR-10 datasets. Additionally we analyze how the performance of the proposed algorithm depends on the dataset size and data augmentation.