Metraipolis › archive › dc_ca43494b51ef

MIND the Gap: A Geographic Implicit Neural Representation with Adjustable Spatial Scale — arXiv:2609.25454

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.25454
License (as archived)
CC0-1.0
Archived
2026-10-08T15:08:15.141Z
SHA-256 of the archived text
3276d9bfdf31e5cad7466fe492476776199c34b11011f21e9c2c66ed05a74a04
Stable URL of this page
https://metraipolis.com/documento/dc_ca43494b51ef

How to cite

MIND the Gap: A Geographic Implicit Neural Representation with Adjustable Spatial Scale — arXiv:2609.25454 — arXiv (en), https://arxiv.org/abs/2609.25454 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ca43494b51ef (Metraipolis archive dc_ca43494b51ef, sha256 3276d9bfdf31e5cad7466fe492476776199c34b11011f21e9c2c66ed05a74a04)
BibTeX
@misc{dcca43494b51ef,
  title = {MIND the Gap: A Geographic Implicit Neural Representation with Adjustable Spatial Scale — arXiv:2609.25454},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.25454},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ca43494b51ef; sha256 3276d9bfdf31e5cad7466fe492476776199c34b11011f21e9c2c66ed05a74a04},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.25454. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Geographic measurements are often sparse, leaving large areas without labels for the quantities we want to map. Geographic implicit neural representations (INRs) provide coordinate-based embeddings that can be combined with sparse labels to predict at unsampled locations without satellite imagery at inference. Yet existing INRs are largely evaluated with random holdouts, leaving their ability to generalize across larger geographic gaps unclear. We introduce Matryoshka Implicit Neural Distillation (MIND), a geographic INR whose spatial granularity can be adjusted after training. MIND distills several pretrained geospatial models using nested supervision at increasing embedding dimensions, dividing the representation into contiguous chunks. Early chunks capture broad spatial patterns, while later chunks add increasingly local variation. Downstream models can retain only the leading chunks or use our Chunked Penalty to reduce reliance on later chunks without retraining the INR. We also introduce CoordBench, comprising $52$ datasets and $78$ targets with both random and regional holdouts at multiple spatial scales. Across CoordBench, fine-scale features help most when labels are nearby, while smoother representations generalize better across larger geographic gaps. MIND with the Chunked Penalty achieves the highest aggregate regression and classification performance among tested INRs and the highest overall performance under regional holdout. These results show that geographic representations should be evaluated and adapted according to the spatial separation between labeled and prediction locations.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-23 in cs.LG).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.25454 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T14:37:29.531Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).