Weakly Supervised Learning For Geospatial Data
Weakly Supervised Learning for Geospatial Data is a machine learning paradigm that leverages incomplete, imprecise, or noisy labels—such as image-level annotations, bounding boxes, or proxy variables—to train models on spatially referenced data, thereby reducing the reliance on densely annotated ground truth while accounting for spatial autocorrelation, scale, and heterogeneity inherent in geospatial applications.
16
Publications
1.3K
Citations
74
Authors
39
Institutions
Publications per year
2017–2026
12
Parents
16
74
2
h-index
2
h-index
2
h-index
2
h-index
2
h-index
1–5 of 74
39
University of Canberra
Canberra, Australia
2 h-index
4 publications
73 citations

UNSW Sydney
Sydney, Australia
2 h-index
4 publications
73 citations

Northwestern Polytechnical University
Xi'an, China
2 h-index
8 publications
3K citations
Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR)
Cologne, Germany
2 h-index
2 publications
85 citations
Meta (Israel)
Tel Aviv, Israel
1 h-index
1 publications
34 citations
RIKEN Center for Advanced Intelligence Project
Tokyo, Japan
1 h-index
1 publications
75 citations

Chinese Academy of Sciences
Beijing, China
1 h-index
3 publications
75 citations
The University of Tokyo
Tokyo, Japan
1 h-index
1 publications
75 citations
1–8 of 39
2 publications · 94 citations
2
h-index
2 publications · 750 citations
2
h-index
2 publications · 43 citations
2
h-index
3 publications · 115 citations
2
h-index
1 publications · 57 citations
1
h-index
1–5