Robust Learning with Jacobian Regularization
This work addresses the need for reliable and stable machine learning systems, particularly in applications sensitive to data corruption, though it is incremental as it builds on existing regularization techniques.
The paper tackled the problem of ensuring neural network robustness against input perturbations by developing a computationally efficient Jacobian regularization method, which improved robustness against random and adversarial perturbations without significantly harming generalization on clean data.
Design of reliable systems must guarantee stability against input perturbations. In machine learning, such guarantee entails preventing overfitting and ensuring robustness of models against corruption of input data. In order to maximize stability, we analyze and develop a computationally efficient implementation of Jacobian regularization that increases classification margins of neural networks. The stabilizing effect of the Jacobian regularizer leads to significant improvements in robustness, as measured against both random and adversarial input perturbations, without severely degrading generalization properties on clean data.