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 is a task of answering a sequence of questions grounded in animage using the previous dialog history as context. In this paper, we study howto address two fundamental challenges for this task: (1) reasoning overunderlying semantic structures among dialog rounds and (2) identifying severalappropriate answers to the given question. To address these challenges, wepropose a Sparse Graph Learning (SGL) method to formulate visual dialog as agraph structure learning task. SGL infers inherently sparse dialog structuresby incorporating binary and score edges and leveraging a new structural lossfunction. Next, we introduce a Knowledge Transfer (KT) method that extracts theanswer predictions from the teacher model and uses them as pseudo labels. Wepropose KT to remedy the shortcomings of single ground-truth labels, whichseverely limit the ability of a model to obtain multiple reasonable answers. Asa result, our proposed model significantly improves reasoning capabilitycompared to baseline methods and outperforms the state-of-the-art approaches onthe VisDial v1.0 dataset. The source code is available athttps://github.com/gicheonkang/SGLKT-VisDial.

Reasoning Visual Dialog with Sparse Graph Learning and Knowledge Transfer