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

People typically learn through exposure to visual concepts associated withlinguistic descriptions. For instance, teaching visual object categories tochildren is often accompanied by descriptions in text or speech. In a machinelearning context, these observations motivates us to ask whether this learningprocess could be computationally modeled to learn visual classifiers. Morespecifically, the main question of this work is how to utilize purely textualdescription of visual classes with no training images, to learn explicit visualclassifiers for them. We propose and investigate two baseline formulations,based on regression and domain transfer, that predict a linear classifier.Then, we propose a new constrained optimization formulation that combines aregression function and a knowledge transfer function with additionalconstraints to predict the parameters of a linear classifier. We also propose ageneric kernelized models where a kernel classifier is predicted in the formdefined by the representer theorem. The kernelized models allow defining andutilizing any two RKHS (Reproducing Kernel Hilbert Space) kernel functions inthe visual space and text space, respectively. We finally propose a kernelfunction between unstructured text descriptions that builds on distributionalsemantics, which shows an advantage in our setting and could be useful forother applications. We applied all the studied models to predict visualclassifiers on two fine-grained and challenging categorization datasets (CUBirds and Flower Datasets), and the results indicate successful predictions ofour final model over several baselines that we designed.

Write a Classifier: Predicting Visual Classifiers from Unstructured Text