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

Understanding spatial relations is essential for intelligent agents to actand communicate in the physical world. Relative directions are spatialrelations that describe the relative positions of target objects with regard tothe intrinsic orientation of reference objects. Grounding relative directionsis more difficult than grounding absolute directions because it not onlyrequires a model to detect objects in the image and to identify spatialrelation based on this information, but it also needs to recognize theorientation of objects and integrate this information into the reasoningprocess. We investigate the challenging problem of grounding relativedirections with end-to-end neural networks. To this end, we provide GRiD-3D, anovel dataset that features relative directions and complements existing visualquestion answering (VQA) datasets, such as CLEVR, that involve only absolutedirections. We also provide baselines for the dataset with two establishedend-to-end VQA models. Experimental evaluations show that answering questionson relative directions is feasible when questions in the dataset simulate thenecessary subtasks for grounding relative directions. We discover that thosesubtasks are learned in an order that reflects the steps of an intuitivepipeline for processing relative directions.

What is Right for Me is Not Yet Right for You: A Dataset for Grounding Relative Directions via Multi-Task Learning