Reconstructing Pruned Filters using Cheap Spatial Transformations
This addresses computational efficiency for deep learning practitioners, though it appears incremental as it builds on existing pruning approaches.
The paper tackles the problem of parameter inefficiency in convolutional neural networks by replacing pruned filters with cheap spatial transformations of remaining filters, achieving comparable or improved performance to state-of-the-art pruning models on CIFAR-10 and ImageNet-1K datasets.
We present an efficient alternative to the convolutional layer using cheap spatial transformations. This construction exploits an inherent spatial redundancy of the learned convolutional filters to enable a much greater parameter efficiency, while maintaining the top-end accuracy of their dense counter-parts. Training these networks is modelled as a generalised pruning problem, whereby the pruned filters are replaced with cheap transformations from the set of non-pruned filters. We provide an efficient implementation of the proposed layer, followed by two natural extensions to avoid excessive feature compression and to improve the expressivity of the transformed features. We show that these networks can achieve comparable or improved performance to state-of-the-art pruning models across both the CIFAR-10 and ImageNet-1K datasets.