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

Given an object of interest, visual navigation aims to reach the object'slocation based on a sequence of partial observations. To this end, an agentneeds to 1) learn a piece of certain knowledge about the relations of objectcategories in the world during training and 2) look for the target object basedon the pre-learned object category relations and its moving trajectory in thecurrent unseen environment. In this paper, we propose a Category Relation Graph(CRG) to learn the knowledge of object category layout relations and aTemporal-Spatial-Region (TSR) attention architecture to perceive the long-termspatial-temporal dependencies of objects helping the navigation. We learn priorknowledge of object layout, establishing a category relationship graph todeduce the positions of specific objects. Subsequently, we introduced TSR tocapture the relationships of objects in temporal, spatial, and regions withinthe observation trajectories. Specifically, we propose a Temporal attentionmodule (T) to model the temporal structure of the observation sequence, whichimplicitly encodes the historical moving or trajectory information. Then, aSpatial attention module (S) is used to uncover the spatial context of thecurrent observation objects based on the category relation graph and pastobservations. Last, a Region attention module (R) shifts the attention to thetarget-relevant region. Based on the visual representation extracted by ourmethod, the agent can better perceive the environment and easily learn superiornavigation policy. Experiments on AI2-THOR demonstrate our CRG-TSR methodsignificantly outperforms existing methods regarding both effectiveness andefficiency. The code has been included in the supplementary material and willbe publicly available.

Building Category Graphs Representation with Spatial and Temporal Attention for Visual Navigation