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

Service robots should be able to interact naturally with non-expert humanusers, not only to help them in various tasks but also to receive guidance inorder to resolve ambiguities that might be present in the instruction. Weconsider the task of visual grounding, where the agent segments an object froma crowded scene given a natural language description. Modern holisticapproaches to visual grounding usually ignore language structure and struggleto cover generic domains, therefore relying heavily on large datasets.Additionally, their transfer performance in RGB-D datasets suffers due to highvisual discrepancy between the benchmark and the target domains. Modularapproaches marry learning with domain modeling and exploit the compositionalnature of language to decouple visual representation from language parsing, buteither rely on external parsers or are trained in an end-to-end fashion due tothe lack of strong supervision. In this work, we seek to tackle theselimitations by introducing a fully decoupled modular framework forcompositional visual grounding of entities, attributes, and spatial relations.We exploit rich scene graph annotations generated in a synthetic domain andtrain each module independently. Our approach is evaluated both in simulationand in two real RGB-D scene datasets. Experimental results show that thedecoupled nature of our framework allows for easy integration with domainadaptation approaches for Sim-To-Real visual recognition, offering adata-efficient, robust, and interpretable solution to visual grounding inrobotic applications.

Sim-To-Real Transfer of Visual Grounding for Human-Aided Ambiguity Resolution