DeepConsensus: using the consensus of features from multiple layers to attain robust image classification
read the original abstract
We consider a classifier whose test set is exposed to various perturbations that are not present in the training set. These test samples still contain enough features to map them to the same class as their unperturbed counterpart. Current architectures exhibit rapid degradation of accuracy when trained on standard datasets but then used to classify perturbed samples of that data. To address this, we present a novel architecture named DeepConsensus that significantly improves generalization to these test-time perturbations. Our key insight is that deep neural networks should directly consider summaries of low and high level features when making classifications. Existing convolutional neural networks can be augmented with DeepConsensus, leading to improved resistance against large and small perturbations on MNIST, EMNIST, FashionMNIST, CIFAR10 and SVHN datasets.
This paper has not been read by Pith yet.
Forward citations
Cited by 1 Pith paper
-
Variations on the Chebyshev-Lagrange Activation Function
Chebyshev-Lagrange activations with linear extrapolation match or exceed ReLU/tanh performance in residual networks on image and vector classification tasks.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.