Convolutional neural networks (CNNs) are widely used for classification problems. However, they often require large amounts of computation and memory which are not readily available in resource constrained systems. Pruning unimportant parameters from CNNs to reduce these requirements has been a subject of intensive research in recent years. However, novel approaches in pruning signals are sometimes difficult to compare against each other. We propose a taxonomy that classifies pruning signals based on four mostly-orthogonal components of the signal. We also empirically evaluate 396 pruning signals including existing ones, and new signals constructed from the components of existing signals. We find that some of our newly constructed signals outperform the best existing pruning signals.
Deep Neural Networks are highly over-parameterized and the size of the neural networks can be reduced significantly after training without any decrease in performance. One can clearly see this phenomenon in a wide range of architectures trained for various problems. Weight/channel pruning, distillation, quantization, matrix factorization are some of the main methods one can use to remove the redundancy to come up with smaller and faster models. This work starts with a short informative chapter, where we motivate the pruning idea and provide the necessary notation. In the second chapter, we compare various saliency scores in the context of parameter pruning. Using the insights obtained from this comparison and stating the problems it brings we motivate why pruning units instead of the individual parameters might be a better idea. We propose some set of definitions to quantify and analyze units that don't learn and create any useful information. We propose an efficient way for detecting dead units and use it to select which units to prune. We get 5x model size reduction through unit-wise pruning on MNIST.
Existing methods for reducing the computational burden of neural networks at run-time, such as parameter pruning or dynamic computational path selection, focus solely on improving computational efficiency during inference. On the other hand, in this work, we propose a novel method which reduces the memory footprint and number of computing operations required for training and inference. Our framework efficiently integrates pruning as part of the training procedure by exploring and tracking the relative importance of convolutional channels. At each training step, we select only a subset of highly salient channels to execute according to the combinatorial upper confidence bound algorithm, and run a forward and backward pass only on these activated channels, hence learning their parameters. Consequently, we enable the efficient discovery of compact models. We validate our approach empirically on state-of-the-art CNNs - VGGNet, ResNet and DenseNet, and on several image classification datasets. Results demonstrate our framework for dynamic channel execution reduces computational cost up to 4x and parameter count up to 9x, thus reducing the memory and computational demands for discovering and training compact neural network models.
We propose a new formulation for pruning convolutional kernels in neural networks to enable efficient inference. We interleave greedy criteria-based pruning with fine-tuning by backpropagation - a computationally efficient procedure that maintains good generalization in the pruned network. We propose a new criterion based on Taylor expansion that approximates the change in the cost function induced by pruning network parameters. We focus on transfer learning, where large pretrained networks are adapted to specialized tasks. The proposed criterion demonstrates superior performance compared to other criteria, e.g. the norm of kernel weights or feature map activation, for pruning large CNNs after adaptation to fine-grained classification tasks (Birds-200 and Flowers-102) relaying only on the first order gradient information. We also show that pruning can lead to more than 10x theoretical (5x practical) reduction in adapted 3D-convolutional filters with a small drop in accuracy in a recurrent gesture classifier. Finally, we show results for the large-scale ImageNet dataset to emphasize the flexibility of our approach.
Predicting human fixations from images has recently seen large improvements by leveraging deep representations which were pretrained for object recognition. However, as we show in this paper, these networks are highly overparameterized for the task of fixation prediction. We first present a simple yet principled greedy pruning method which we call Fisher pruning. Through a combination of knowledge distillation and Fisher pruning, we obtain much more runtime-efficient architectures for saliency prediction, achieving a 10x speedup for the same AUC performance as a state of the art network on the CAT2000 dataset. Speeding up single-image gaze prediction is important for many real-world applications, but it is also a crucial step in the development of video saliency models, where the amount of data to be processed is substantially larger.