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

Visual Dialog aims to answer multi-round, interactive questions based on thedialog history and image content. Existing methods either consider answerranking and generating individually or only weakly capture the relation acrossthe two tasks implicitly by two separate models. The research on a universalframework that jointly learns to rank and generate answers in a single model isseldom explored. In this paper, we propose a contrastive learning-basedframework UTC to unify and facilitate both discriminative and generative tasksin visual dialog with a single model. Specifically, considering the inherentlimitation of the previous learning paradigm, we devise two inter-taskcontrastive losses i.e., context contrastive loss and answer contrastive lossto make the discriminative and generative tasks mutually reinforce each other.These two complementary contrastive losses exploit dialog context and targetanswer as anchor points to provide representation learning signals fromdifferent perspectives. We evaluate our proposed UTC on the VisDial v1.0dataset, where our method outperforms the state-of-the-art on bothdiscriminative and generative tasks and surpasses previous state-of-the-artgenerative methods by more than 2 absolute points on Recall@1.

UTC: A Unified Transformer with Inter-Task Contrastive Learning for Visual Dialog