ChoiceNet: CNN learning through choice of multiple feature map representations
This work addresses efficiency and gradient issues in deep learning for computer vision, but it appears incremental as it builds on existing skip connection ideas.
The authors tackled the problem of vanishing gradients and parameter efficiency in CNNs by introducing ChoiceNet, a highly connected architecture with skip connections and channelwise concatenations, achieving competitive performance on multiple object recognition and semantic segmentation datasets.
We introduce a new architecture called ChoiceNet where each layer of the network is highly connected with skip connections and channelwise concatenations. This enables the network to alleviate the problem of vanishing gradients, reduces the number of parameters without sacrificing performance, and encourages feature reuse. We evaluate our proposed architecture on three benchmark datasets for object recognition tasks (ImageNet, CIFAR- 10, CIFAR-100, SVHN) and on a semantic segmentation dataset (CamVid).