Reducing the Uncertainty in Determining the Origin of Natural Commodities via Earth Embeddings

Abstract

Verifying where commodities like timber were harvested is a forensic task, needed to detect illegal trade due to illegal deforestation or trade sanctions. Existing Gaussian-process models of wood chemistry give a Bayesian posterior over candidate harvest locations, but their credible regions in general do not deliver the coverage they promise (0.82-0.91 at the nominal 95\%). In this work we bring in a second, independent source of evidence: a prior over harvest locations learned from satellite (AlphaEarth) embeddings. We fit one classifier per genus that predicts presence from an embedding vector, supervised by occurrence records, and normalize its scores into a prior over the candidate grid. For such a prior to act on the prediction sets rather than merely mask cells, we build a cross-conformal predictor on the \emph{posterior cell mass}, so the prior moves every p-value while the coverage guarantee survives any misspecification of it. On four European genera, the embedding prior shrinks regions by up to 42\% compared to a uniform prior at no cost in coverage, which holds at 0.953-0.968 against a nominal 0.95.

Recognizability of Being Point Determining
M. H. Shirdareh Haghighi, A. Namazi, Z Rahimi +1
arXiv28 Sep 2026PreprintarXiv licenceFull textCombinatorics & discrete mathematicsarXiv
Determining norm-Euclidean complex cubic fields
Gustav Kjærbye Bagger, Andrew R. Booker, Tim Trudgian
arXiv27 Oct 2026PreprintarXiv licenceFull textAlgebra & number theoryarXiv
Determining the Rearrangement Number
Vinicius de Oliveira Rodrigues
arXiv6 Sep 2026PreprintCC BY 4.0Full textCombinatorics & discrete mathematics Logic & foundationsarXiv
Determining Schrodinger Potentials on Riemannian Manifolds
Lu Chen, Yan Jiang, Hongyu Liu +1
arXiv8 Oct 2026PreprintCC BY 4.0Full textAnalysis & PDEsarXiv
Functions determining few direction: the final missing case
Peter Sziklai
arXiv20 Oct 2026PreprintCC BY 4.0Full textCombinatorics & discrete mathematicsarXiv

Cite

BibTeX RIS

BibTeX

@misc{jimenez2026reducing,
  title = {{Reducing the Uncertainty in Determining the Origin of Natural Commodities via Earth Embeddings}},
  author = {Sebastián Jiménez and Willem Waegeman and Victor Deklerck and Thomas Mortier},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.35393},
  eprint = {2610.35393},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.35393},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - Reducing the Uncertainty in Determining the Origin of Natural Commodities via Earth Embeddings
AU  - Sebastián Jiménez
AU  - Willem Waegeman
AU  - Victor Deklerck
AU  - Thomas Mortier
PY  - 2026
DA  - 2026/10/26
PB  - arXiv
DO  - 10.48550/arxiv.2610.35393
UR  - https://arxiv.org/abs/2610.35393
AB  - Verifying where commodities like timber were harvested is a forensic task, needed to detect illegal trade due to illegal deforestation or trade sanctions. Existing Gaussian-process models of wood chemistry give a Bayesian posterior over candidate harvest locations, but their credible regions in general do not deliver the coverage they promise (0.82-0.91 at the nominal 95\%). In this work we bring in a second, independent source of evidence: a prior over harvest locations learned from satellite (AlphaEarth) embeddings. We fit one classifier per genus that predicts presence from an embedding vector, supervised by occurrence records, and normalize its scores into a prior over the candidate grid. For such a prior to act on the prediction sets rather than merely mask cells, we build a cross-conformal predictor on the \emph{posterior cell mass}, so the prior moves every p-value while the coverage guarantee survives any misspecification of it. On four European genera, the embedding prior shrinks regions by up to 42\% compared to a uniform prior at no cost in coverage, which holds at 0.953-0.968 against a nominal 0.95.
ER  -

Abstract

Related

Cite