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 neural language model can be conditioned into generating descriptions forimages by providing visual information apart from the sentence prefix. Thisvisual information can be included into the language model through differentpoints of entry resulting in different neural architectures. We identify fourmain architectures which we call init-inject, pre-inject, par-inject, andmerge.  We analyse these four architectures and conclude that the best performing oneis init-inject, which is when the visual information is injected into theinitial state of the recurrent neural network. We confirm this using bothautomatic evaluation measures and human annotation.  We then analyse how much influence the images have on each architecture. Thisis done by measuring how different the output probabilities of a model are whena partial sentence is combined with a completely different image from the oneit is meant to be combined with. We find that init-inject tends to quicklybecome less influenced by the image as more words are generated. A differentarchitecture called merge, which is when the visual information is merged withthe recurrent neural network's hidden state vector prior to output, losesvisual influence much more slowly, suggesting that it would work better forgenerating longer sentences.  We also observe that the merge architecture can have its recurrent neuralnetwork pre-trained in a text-only language model (transfer learning) ratherthan be initialised randomly as usual. This results in even better performancethan the other architectures, provided that the source language model is nottoo good at language modelling or it will overspecialise and be less effectiveat image description generation.  Our work opens up new avenues of research in neural architectures,explainable AI, and transfer learning.

On Architectures for Including Visual Information in Neural Language Models for Image Description