2021
DOI: 10.48550/arxiv.2111.12872
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Less is More: Generating Grounded Navigation Instructions from Landmarks

Abstract: We study the automatic generation of navigation instructions from 360 • images captured on indoor routes. Existing generators suffer from poor visual grounding, causing them to rely on language priors and hallucinate objects. Our MARKY-MT5 system addresses this by focusing on visual landmarks; it comprises a first stage landmark detector and a second stage generator-a multimodal, multilingual, multitask encoder-decoder. To train it, we bootstrap grounded landmark annotations on top of the Room-across-Room (RxR… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...

Citation Types

0
0
0

Publication Types

Select...

Relationship

0
0

Authors

Journals

citations
Cited by 0 publications
references
References 51 publications
0
0
0
Order By: Relevance

No citations

Set email alert for when this publication receives citations?