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

Top-down information plays a central role in human perception, but playsrelatively little role in many current state-of-the-art deep networks, such asConvolutional Neural Networks (CNNs). This work seeks to explore a path bywhich top-down information can have a direct impact within current deepnetworks. We explore this path by learning and using "generators" correspondingto the network internal effects of three types of transformation (each arestriction of a general affine transformation): rotation, scaling, andtranslation. We demonstrate how these learned generators can be used totransfer top-down information to novel settings, as mediated by the "featureflows" that the transformations (and the associated generators) correspond toinside the network. Specifically, we explore three aspects: 1) using generatorsas part of a method for synthesizing transformed images --- given a previouslyunseen image, produce versions of that image corresponding to one or morespecified transformations, 2) "zero-shot learning" --- when provided with afeature flow corresponding to the effect of a transformation of unknown amount,leverage learned generators as part of a method by which to perform an accuratecategorization of the amount of transformation, even for amounts never observedduring training, and 3) (inside-CNN) "data augmentation" --- improve theclassification performance of an existing network by using the learnedgenerators to directly provide additional training "inside the CNN".

What Happened to My Dog in That Network: Unraveling Top-down Generators in Convolutional Neural Networks