English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

A recent line of work showed that various forms of convolutional kernelmethods can be competitive with standard supervised deep convolutional networkson datasets like CIFAR-10, obtaining accuracies in the range of 87-90% whilebeing more amenable to theoretical analysis. In this work, we highlight theimportance of a data-dependent feature extraction step that is key to theobtain good performance in convolutional kernel methods. This step typicallycorresponds to a whitened dictionary of patches, and gives rise to adata-driven convolutional kernel methods. We extensively study its effect,demonstrating it is the key ingredient for high performance of these methods.Specifically, we show that one of the simplest instances of such kernelmethods, based on a single layer of image patches followed by a linearclassifier is already obtaining classification accuracies on CIFAR-10 in thesame range as previous more sophisticated convolutional kernel methods. Wescale this method to the challenging ImageNet dataset, showing such a simpleapproach can exceed all existing non-learned representation methods. This is anew baseline for object recognition without representation learning methods,that initiates the investigation of convolutional kernel models on ImageNet. Weconduct experiments to analyze the dictionary that we used, our ablationsshowing they exhibit low-dimensional properties.

The Unreasonable Effectiveness of Patches in Deep Convolutional Kernels Methods