Convolutional Neural Fabrics

TL;DR

This paper introduces a neural network 'fabric' that embeds a vast space of architectures, allowing for flexible model selection and ensembling, with competitive results on image classification and segmentation tasks.

Contribution

The paper proposes a novel neural fabric structure that encapsulates many architectures within a single model, enabling efficient training and ensemble capabilities.

Findings

Competitive performance on MNIST and CIFAR10

Effective ensembling of multiple architectures within one fabric

Linear scaling of training cost with fabric size

Abstract

Despite the success of CNNs, selecting the optimal architecture for a given task remains an open problem. Instead of aiming to select a single optimal architecture, we propose a "fabric" that embeds an exponentially large number of architectures. The fabric consists of a 3D trellis that connects response maps at different layers, scales, and channels with a sparse homogeneous local connectivity pattern. The only hyper-parameters of a fabric are the number of channels and layers. While individual architectures can be recovered as paths, the fabric can in addition ensemble all embedded architectures together, sharing their weights where their paths overlap. Parameters can be learned using standard methods based on back-propagation, at a cost that scales linearly in the fabric size. We present benchmark results competitive with the state of the art for image classification on MNIST and CIFAR10, and for semantic segmentation on the Part Labels dataset.