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

In the dataset of image captioning, each image is aligned with severaldescriptions. Despite the fact that the quality of these descriptions varies,existing captioning models treat them equally in the training process. In thispaper, we propose a new control signal of sentence quality, which is taken asan additional input to the captioning model. By integrating the control signalinformation, captioning models are aware of the quality level of the targetsentences and handle them differently. Moreover, we propose a novelreinforcement training method specially designed for the control signal ofsentence quality: Quality-oriented Self-Annotated Training (Q-SAT). Extensiveexperiments on MSCOCO dataset show that without extra information from groundtruth captions, models controlled by the highest quality level outperformbaseline models on accuracy-based evaluation metrics, which validates theeffectiveness of our proposed methods.

Improving Image Captioning with Control Signal of Sentence Quality