Compression-aware Training of Deep Networks
This addresses the issue of high computational and memory costs in deep learning for practitioners, though it is incremental as it builds on existing compression strategies.
The paper tackles the problem of deep neural networks being computationally and memory expensive by proposing a compression-aware training method that incorporates a low-rank regularizer during training, resulting in models that are more compact and at least as effective as state-of-the-art compression techniques.
In recent years, great progress has been made in a variety of application domains thanks to the development of increasingly deeper neural networks. Unfortunately, the huge number of units of these networks makes them expensive both computationally and memory-wise. To overcome this, exploiting the fact that deep networks are over-parametrized, several compression strategies have been proposed. These methods, however, typically start from a network that has been trained in a standard manner, without considering such a future compression. In this paper, we propose to explicitly account for compression in the training process. To this end, we introduce a regularizer that encourages the parameter matrix of each layer to have low rank during training. We show that accounting for compression during training allows us to learn much more compact, yet at least as effective, models than state-of-the-art compression techniques.