Sifting out the features by pruning: Are convolutional networks the winning lottery ticket of fully connected ones?
This work addresses the problem of understanding inductive biases in neural network pruning for researchers in machine learning, but it is incremental as it builds on existing pruning concepts.
The study investigated whether pruning fully connected networks (FCNs) on visual tasks reveals sub-networks with local connectivity patterns similar to convolutional networks (CNNs), showing that these 'winning lottery tickets' exhibit key CNN features without specifying performance metrics.
Pruning methods can considerably reduce the size of artificial neural networks without harming their performance. In some cases, they can even uncover sub-networks that, when trained in isolation, match or surpass the test accuracy of their dense counterparts. Here we study the inductive bias that pruning imprints in such "winning lottery tickets". Focusing on visual tasks, we analyze the architecture resulting from iterative magnitude pruning of a simple fully connected network (FCN). We show that the surviving node connectivity is local in input space, and organized in patterns reminiscent of the ones found in convolutional networks (CNN). We investigate the role played by data and tasks in shaping the architecture of pruned sub-networks. Our results show that the winning lottery tickets of FCNs display the key features of CNNs. The ability of such automatic network-simplifying procedure to recover the key features "hand-crafted" in the design of CNNs suggests interesting applications to other datasets and tasks, in order to discover new and efficient architectural inductive biases.