The Effect of Data Dimensionality on Neural Network Prunability

Practitioners prune neural networks for efficiency gains and generalization improvements, but few scrutinize the factors determining the <PRE_TAG>prunability</POST_TAG> of a neural network the maximum fraction of weights that pruning can remove without compromising the model's test accuracy. In this work, we study the properties of input data that may contribute to the <PRE_TAG>prunability</POST_TAG> of a neural network. For high dimensional <PRE_TAG>input data</POST_TAG> such as images, text, and audio, the manifold hypothesis suggests that these high dimensional inputs approximately lie on or near a significantly lower dimensional manifold. Prior work demonstrates that the underlying <PRE_TAG>low dimensional structure</POST_TAG> of the input data may affect the <PRE_TAG>sample efficiency</POST_TAG> of learning. In this paper, we investigate whether the low dimensional structure of the input data affects the <PRE_TAG>prunability</POST_TAG> of a neural network.