What's in an Earth Embedding? An Explainability Analysis of Location Encoders

• URL: http://arxiv.org/abs/2606.24997v1
• Date: Tue, 23 Jun 2026 16:01:17 GMT
• Title: What's in an Earth Embedding? An Explainability Analysis of Location Encoders
• Abstract summary: Geographic embeddings (INRs) learn to map any coordinate on Earth to a location.<n>We analyze these embeddings into human-interpretable representations.<n>We show that location embeddings can be decomposed into human-interpretable representations.
• Score: 6.103957378939232
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Geographic implicit neural representations (INRs) learn to map any coordinate on Earth to a location embedding, implicitly encoding geospatial data into the weights of a neural network. Location embeddings are widely used off the shelf as general-purpose geospatial representations, yet users lack principled tools to audit what geographic or semantic information these embeddings capture. In this work, we analyze the information content of geographic INRs through their location embeddings. We decompose these embeddings into human-interpretable features$\unicode{x2014}$namely, (i) sparse latent concepts, (ii) natural language concepts, and (iii) visual features. The latent concept embeddings are learned using sparse autoencoders. To recover natural language concepts, we apply sparse linear concept embeddings (SpLiCE) over a predefined geospatial dictionary. Finally, visual features are extracted using saliency maps derived from CLIP Surgery. We show that location embeddings can be decomposed into human-interpretable representations while retaining high reconstruction capability, revealing interpretable geographic structures such as forests, deserts, and urban features. Across methods, sparse decompositions expose systematic differences in encoded information, ranging from urban structures to broader biome and climate signals, and pretraining-space saliency maps further highlight complementary features such as roads and landmarks. We hope this work provides a first step toward interpretable geospatial representations.

Related papers

• Tessellating The Earth [16.078192224421084]
  Tessellating the Earth (TTE) is a location encoder built from learnable Spherical Voronoi partitions.<n>TTE sets a new state of the art for location encoders across a suite of geospatial classification and regression tasks.
  arXiv  Detail & Related papers  (2026-06-25T19:58:13Z)
• Spatial Representation Learning Beyond Pixels: Unifying Raster Data and Vector Semantics for Human-Centric Geospatial Foundation Models [12.200359585933766]
  Recent advances in self-supervised learning have given rise to Earth Observation Foundation Models (EOFMs)<n>This paper calls for a paradigm shift toward joint Spatial Representation Learning (SRL) in an unified embedding space that integrate perception with vector-based reasoning.<n>We contend that such integration is essential for developing next-generation geospatial AI systems capable of more accurate, interpretable, and grounded understanding of the Earth.
  arXiv  Detail & Related papers  (2026-06-01T15:21:58Z)
• Geospatial Mechanistic Interpretability of Large Language Models [6.0272491755196045]
  Large Language Models (LLMs) have demonstrated unprecedented capabilities across various natural language processing tasks.<n>Our aim is to advance our understanding of the internal representations that these complex models generate while processing geographical information.
  arXiv  Detail & Related papers  (2025-05-06T09:40:06Z)
• FindAnything: Open-Vocabulary and Object-Centric Mapping for Robot Exploration in Any Environment [16.987872206495897]
  FindAnything is an open-world mapping framework that incorporates vision-language information into dense volumetric submaps.<n>Our system is the first of its kind to be deployed on resource-constrained devices, such as MAVs.
  arXiv  Detail & Related papers  (2025-04-11T15:12:05Z)
• MapExplorer: New Content Generation from Low-Dimensional Visualizations [60.02149343347818]
  Low-dimensional visualizations, or "projection maps," are widely used to interpret large-scale and complex datasets.<n>These visualizations not only aid in understanding existing knowledge spaces but also implicitly guide exploration into unknown areas.<n>We introduce MapExplorer, a novel knowledge discovery task that translates coordinates within any projection map into coherent, contextually aligned textual content.
  arXiv  Detail & Related papers  (2024-12-24T20:16:13Z)
• Mapping High-level Semantic Regions in Indoor Environments without Object Recognition [50.624970503498226]
  The present work proposes a method for semantic region mapping via embodied navigation in indoor environments. To enable region identification, the method uses a vision-to-language model to provide scene information for mapping. By projecting egocentric scene understanding into the global frame, the proposed method generates a semantic map as a distribution over possible region labels at each location.
  arXiv  Detail & Related papers  (2024-03-11T18:09:50Z)
• SatCLIP: Global, General-Purpose Location Embeddings with Satellite Imagery [22.716322265391852]
  We introduce Satellite Contrastive Location-Image Pretraining (SatCLIP) SatCLIP learns an implicit representation of locations by matching CNN and ViT inferred visual patterns of openly available satellite imagery with their geographic coordinates. In experiments, we use SatCLIP embeddings to improve prediction performance on nine diverse location-dependent tasks including temperature prediction, animal recognition, and population density estimation.
  arXiv  Detail & Related papers  (2023-11-28T19:14:40Z)
• Automated mapping of virtual environments with visual predictive coding [0.9591674293850556]
  We introduce a framework in which an agent navigates a virtual environment while engaging in visual predictive coding. While learning a next image prediction task, the agent automatically constructs an internal representation of the environment that quantitatively reflects distances. The internal map enables the agent to pinpoint its location relative to landmarks using only visual information.
  arXiv  Detail & Related papers  (2023-08-20T22:29:16Z)
• GeoGLUE: A GeoGraphic Language Understanding Evaluation Benchmark [56.08664336835741]
  We propose a GeoGraphic Language Understanding Evaluation benchmark, named GeoGLUE. We collect data from open-released geographic resources and introduce six natural language understanding tasks. We pro vide evaluation experiments and analysis of general baselines, indicating the effectiveness and significance of the GeoGLUE benchmark.
  arXiv  Detail & Related papers  (2023-05-11T03:21:56Z)
• Latent Topology Induction for Understanding Contextualized Representations [84.7918739062235]
  We study the representation space of contextualized embeddings and gain insight into the hidden topology of large language models. We show there exists a network of latent states that summarize linguistic properties of contextualized representations.
  arXiv  Detail & Related papers  (2022-06-03T11:22:48Z)
• PGL: Prior-Guided Local Self-supervised Learning for 3D Medical Image Segmentation [87.50205728818601]
  We propose a PriorGuided Local (PGL) self-supervised model that learns the region-wise local consistency in the latent feature space. Our PGL model learns the distinctive representations of local regions, and hence is able to retain structural information.
  arXiv  Detail & Related papers  (2020-11-25T11:03:11Z)
• SIRI: Spatial Relation Induced Network For Spatial Description Resolution [64.38872296406211]
  We propose a novel relationship induced (SIRI) network for language-guided localization. We show that our method is around 24% better than the state-of-the-art method in terms of accuracy, measured by an 80-pixel radius. Our method also generalizes well on our proposed extended dataset collected using the same settings as Touchdown.
  arXiv  Detail & Related papers  (2020-10-27T14:04:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.