Group Ensemble: Learning an Ensemble of ConvNets in a single ConvNet
This addresses the problem of high computational cost for ensemble methods in deep learning, making them more practical for applications like image recognition, though it is an incremental improvement over existing ensemble techniques.
The paper tackles the computational inefficiency of ensemble learning in deep learning by proposing Group Ensemble Network (GENet), which incorporates an ensemble of ConvNets into a single ConvNet, achieving a 1.83% reduction in top-1 error for ResNeXt-50 on ImageNet.
Ensemble learning is a general technique to improve accuracy in machine learning. However, the heavy computation of a ConvNets ensemble limits its usage in deep learning. In this paper, we present Group Ensemble Network (GENet), an architecture incorporating an ensemble of ConvNets in a single ConvNet. Through a shared-base and multi-head structure, GENet is divided into several groups to make explicit ensemble learning possible in a single ConvNet. Owing to group convolution and the shared-base, GENet can fully leverage the advantage of explicit ensemble learning while retaining the same computation as a single ConvNet. Additionally, we present Group Averaging, Group Wagging and Group Boosting as three different strategies to aggregate these ensemble members. Finally, GENet outperforms larger single networks, standard ensembles of smaller networks, and other recent state-of-the-art methods on CIFAR and ImageNet. Specifically, group ensemble reduces the top-1 error by 1.83% for ResNeXt-50 on ImageNet. We also demonstrate its effectiveness on action recognition and object detection tasks.