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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep Convolutional Neural Networks (CNNs), trained extensively on very largelabeled datasets, learn to recognize inferentially powerful features in theirinput patterns and represent efficiently their objective content. Suchobjectivity of their internal representations enables deep CNNs to readilytransfer and successfully apply these representations to new classificationtasks. Deep CNNs develop their internal representations through a challengingprocess of error backpropagation-based supervised training. In contrast, deepneural networks of the cerebral cortex develop their even more powerfulinternal representations in an unsupervised process, apparently guided at alocal level by contextual information. Implementing such local contextualguidance principles in a single-layer CNN architecture, we propose an efficientalgorithm for developing broad-purpose representations (i.e., representationstransferable to new tasks without additional training) in shallow CNNs trainedon limited-size datasets. A contextually guided CNN (CG-CNN) is trained ongroups of neighboring image patches picked at random image locations in thedataset. Such neighboring patches are likely to have a common context andtherefore are treated for the purposes of training as belonging to the sameclass. Across multiple iterations of such training on different context-sharinggroups of image patches, CNN features that are optimized in one iteration arethen transferred to the next iteration for further optimization, etc. In thisprocess, CNN features acquire higher pluripotency, or inferential utility forany arbitrary classification task, which we quantify as a transfer utility. Inour application to natural images, we find that CG-CNN features show the same,if not higher, transfer utility and classification accuracy as comparabletransferable features in the first CNN layer of the well-known deep networks.

Contextually Guided Convolutional Neural Networks for Learning Most Transferable Representations