Deep Neural Networks with Short Circuits for Improved Gradient Learning
This work addresses the issue of external reliance in deep learning for researchers and practitioners, but it appears incremental as it builds on existing gradient enhancement methods.
The paper tackled the problem of deep neural networks' reliance on external training or computing by proposing a short circuit neural connection approach to enhance gradient learning, resulting in a large margin improvement over baselines on computer vision and natural language processing tasks.
Deep neural networks have achieved great success both in computer vision and natural language processing tasks. However, mostly state-of-art methods highly rely on external training or computing to improve the performance. To alleviate the external reliance, we proposed a gradient enhancement approach, conducted by the short circuit neural connections, to improve the gradient learning of deep neural networks. The proposed short circuit is a unidirectional connection that single back propagates the sensitive from the deep layer to the shallows. Moreover, the short circuit formulates to be a gradient truncation of its crossing layers which can plug into the backbone deep neural networks without introducing external training parameters. Extensive experiments demonstrate deep neural networks with our short circuit gain a large margin over the baselines on both computer vision and natural language processing tasks.