ExpandNets: Linear Over-parameterization to Train Compact Convolutional Networks
arXiv:1811.10495
Abstract
We introduce an approach to training a given compact network. To this end, we leverage over-parameterization, which typically improves both neural network optimization and generalization. Specifically, we propose to expand each linear layer of the compact network into multiple consecutive linear layers, without adding any nonlinearity. As such, the resulting expanded network, or ExpandNet, can be contracted back to the compact one algebraically at inference. In particular, we introduce two convolutional expansion strategies and demonstrate their benefits on several tasks, including image classification, object detection, and semantic segmentation. As evidenced by our experiments, our approach outperforms both training the compact network from scratch and performing knowledge distillation from a teacher. Furthermore, our linear over-parameterization empirically reduces gradient confusion during training and improves the network generalization.
References in corpus (22)
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- Distilling the Knowledge in a Neural Network
- MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
- YOLOv3: An Incremental Improvement
- Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1
- FitNets: Hints for Thin Deep Nets
- Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer
- Pruning Convolutional Neural Networks for Resource Efficient Inference
- Simplifying Graph Convolutional Networks
- A Convergence Theory for Deep Learning via Over-Parameterization
- SNIP: Single-shot Network Pruning based on Connection Sensitivity
- YOLO9000: Better, Faster, Stronger
- Speeding-up Convolutional Neural Networks Using Fine-tuned CP-Decomposition
- Flattened Convolutional Neural Networks for Feedforward Acceleration
- All you need is a good init
- Generalization in Deep Learning
- Learning the Number of Neurons in Deep Networks
- Compression-aware Training of Deep Networks
- PerforatedCNNs: Acceleration through Elimination of Redundant Convolutions
- SqueezeDet: Unified, Small, Low Power Fully Convolutional Neural Networks for Real-Time Object Detection for Autonomous Driving
- Soft Weight-Sharing for Neural Network Compression
- The Impact of Neural Network Overparameterization on Gradient Confusion and Stochastic Gradient Descent