Fine-Pruning: Joint Fine-Tuning and Compression of a Convolutional Network with Bayesian Optimization
This addresses the need for efficient network deployment in specialized visual recognition tasks, but it is incremental as it builds on existing fine-tuning and pruning techniques.
The paper tackles the problem of over-parameterization in fine-tuned deep neural networks for specialized image domains by proposing a method that jointly fine-tunes and compresses networks, overcoming limitations of independent fine-tuning and pruning, with experiments validating the approach on remote sensing and texture datasets.
When approaching a novel visual recognition problem in a specialized image domain, a common strategy is to start with a pre-trained deep neural network and fine-tune it to the specialized domain. If the target domain covers a smaller visual space than the source domain used for pre-training (e.g. ImageNet), the fine-tuned network is likely to be over-parameterized. However, applying network pruning as a post-processing step to reduce the memory requirements has drawbacks: fine-tuning and pruning are performed independently; pruning parameters are set once and cannot adapt over time; and the highly parameterized nature of state-of-the-art pruning methods make it prohibitive to manually search the pruning parameter space for deep networks, leading to coarse approximations. We propose a principled method for jointly fine-tuning and compressing a pre-trained convolutional network that overcomes these limitations. Experiments on two specialized image domains (remote sensing images and describable textures) demonstrate the validity of the proposed approach.