# geospatial

Geospatial data is information tied to locations on Earth, commonly represented using points, lines, polygons, spatial reference systems, imagery, or raster datasets.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## From Geographic Memory to Spatial Generalization in Large Language Models

DevFeed: [From Geographic Memory to Spatial Generalization in Large Language Models](<https://devfeed.tech/articles/from-geographic-memory-to-spatial-generalization-in-large-language-models-145092.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25703>)

Author: Chen Chu, Zhaonan Wang, Cyrus Shahabi, Hengcai Zhang, Feng Lu

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>)

Tags: [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [clustering](<https://devfeed.tech/tags/clustering.md>), [continuous](<https://devfeed.tech/tags/continuous.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [grounding](<https://devfeed.tech/tags/grounding.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [location](<https://devfeed.tech/tags/location.md>), [london](<https://devfeed.tech/tags/london.md>), [spatial](<https://devfeed.tech/tags/spatial.md>)

### AI overview

The study separates geospatial memory--recalling observed location-description pairs--from spatial generalization to locations outside the training observations. It adds continuous location representations from a spatial encoder as tokens in LLMs and evaluates them on newly curated Los Angeles and London benchmarks. Captioning tests observed and unobserved locations, while a separate grounding task measures whether generated pairs help map text to geography. Across multiple LLM families, continuous representations improve captioning at unobserved locations compared with textual coordinates; generated descriptions bring modest but consistent grounding gains. Probing and clustering results are also consistent with spatially organized representations. The evidence is limited to the described benchmarks and assessments.

### Source excerpt

arXiv:2611.25703v1 Announce Type: cross Abstract: Large language models (LLMs) can associate places with geographic descriptions, but recovering an observed association does not establish an ability to generalize across geographic space. We distinguish geospatial memory, the recovery of location--description associations observed during training, from spatial generalization, the prediction of descriptions for locations in unobserved regions. We investigate whether continuous location representations improve this generalization and encourage geographic knowledge to become spatially organized within an LLM. We frame this organization as a latent spatial field: a learned mapping from locations to descriptions connected through spatial dependencies. Building on a continuous spatial encoder, we project location representations into the model's embedding space and supply them as spatial tokens. We evaluate the resulting models on two newly curated geo-text benchmarks for Los Angeles and London using two complementary assessments. Geospatial captioning tests both memory and spatial generalization by generating descriptions at observed locations and at locations beyond the spatial coverage of the training observations. A separate geospatial grounding model tests whether generated location--description pairs provide useful additional supervision for mapping text to geographic space. Across multiple LLM families, continuous location representations improve captioning at unobserved locations relative to textual coordinates, while their generated descriptions yield modest but consistent improvements in downstream grounding. Layer-wise probing and representation clustering provide additional evidence consistent with spatially organized geographic representations. Together, these findings show the value of evaluating spatial generalization separately from memory and of incorporating continuous spatial representations into language models.

## Distilling a Geospatial Foundation Model into Compact Surface-Water Segmenters

DevFeed: [Distilling a Geospatial Foundation Model into Compact Surface-Water Segmenters](<https://devfeed.tech/articles/distilling-a-geospatial-foundation-model-into-compact-surface-water-segmenters-143310.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.20441>)

Author: Fabian Schmalstieg, Karsten Mueller, Wojciech Samek

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Local AI](<https://devfeed.tech/topics/local-ai.md>)

Tags: [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [distilling](<https://devfeed.tech/tags/distilling.md>), [foundation](<https://devfeed.tech/tags/foundation.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [model](<https://devfeed.tech/tags/model.md>), [surface](<https://devfeed.tech/tags/surface.md>), [water](<https://devfeed.tech/tags/water.md>)

### AI overview

A compact student can match a 300-million-parameter Prithvi-EO-2.0 teacher for surface-water segmentation while running on edge hardware. The teacher, fine-tuned on 252 hand-labeled Sen1Floods11 images, labels 2,500 Sentinel-2 images and provides both labels and probability maps for student training. A 0.7-million-parameter EfficientViT-B0 matches teacher water IoU across WorldFloods-v2, STURM-Flood and 97 held-back WorldFloods events; the full pipeline gains 5.7 IoU points on held-back events over training on hand labels alone. On a Jetson Xavier NX, it runs in a 1.6 MB mixed-precision engine at 4.9 ms per 512x512 image. A 12,230-parameter student in full 8-bit integer format stays within 0.5 IoU points of its 32-bit version. Against a fixed water index, the main student gains 11.5 IoU points on held-back events and triggers false alarms on under 10% of water-free windows, versus 57%.

### Source excerpt

arXiv:2609.20441v2 Announce Type: replace-cross Abstract: Geospatial foundation models segment surface water accurately, but their size limits deployment on edge hardware. Therefore, we distill a 300-million-parameter Prithvi-EO-2.0 teacher, fine-tuned on the 252 hand-labeled Sen1Floods11 training images, into compact student models. The teacher labels 2,500 Sentinel-2 images around flood alerts and supervises the students via its labels and probability maps, without manual annotation. A 0.7-million-parameter EfficientViT-B0 student achieves the teacher's accuracy in water intersection over union (IoU) on WorldFloods-v2, STURM-Flood and 97 held-back WorldFloods events. Unlike prior work, we evaluate every model on four test sets from three public datasets and on two new dry-scene sets. Controlled comparisons show that teacher labels achieve human-level annotation quality for student training on the same images. The teacher's probability maps account for much of its value, and the full pipeline adds 5.7 IoU points on the held-back events over training on hand labels alone. All students, with parameter counts ranging from about 1,000 to 32.5 million, perform within 2 IoU points of the teacher on WorldFloods-v2 and STURM-Flood. On a Jetson Xavier NX, the main student runs as a 1.6-megabyte mixed 8-bit integer and 16-bit floating-point engine at 4.9 milliseconds per 512-by-512 image, and a 12,230-parameter student runs in full 8-bit integer format within 0.5 IoU points of its 32-bit version. The main student outperforms a fixed water index by 11.5 IoU points on the held-back events and raises false alarms on under 10\% rather than 57.0\% of water-free windows, as unattended use requires.

## datasette-maplibre is a new geospatial plugin for Datasette

DevFeed: [datasette-maplibre is a new geospatial plugin for Datasette](<https://devfeed.tech/articles/datasette-maplibre-is-a-new-geospatial-plugin-for-datasette-140684.md>)

Original publisher: [Read original article](<https://blog.adafruit.com/2026/11/18/datasette-maplibre-is-a-new-geospatial-plugin-for-datasette/>)

Author: Anne Barela

Published: 2026-11-18T14:25:36Z

Content type: news

Language: en

Sources: [Adafruit Industries](<https://devfeed.tech/sources/adafruit-industries-makers-hackers-artists-designers-and-engineers.md>)

Topics: [geospatial](<https://devfeed.tech/topics/geospatial.md>), [GeoJSON](<https://devfeed.tech/topics/geojson.md>)

Tags: [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [datasette](<https://devfeed.tech/tags/datasette.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [maplibre](<https://devfeed.tech/tags/maplibre.md>), [maps](<https://devfeed.tech/tags/maps.md>), [plugin](<https://devfeed.tech/tags/plugin.md>), [programming](<https://devfeed.tech/tags/programming.md>), [simon-willison](<https://devfeed.tech/tags/simon-willison.md>)

### AI overview

datasette-maplibre 0.1a0 combines clustered latitude/longitude points and GeoJSON shapes on one map, using MapLibre GL with OpenFreeMap. It also previews GeoJSON shapes as SVGs and loads a WebGL preview only when a shape scrolls into view. A demo runs in Datasette Lite, loading the plugin and sample database in a browser through Pyodide and WebAssembly. The excerpt says much of the plugin came from a prompt to Claude that referenced two older Datasette plugins and an OpenFreeMap demo.

### Source excerpt

Simon Willison has released datasette-maplibre 0.1a0 on his blog, the first new geospatial plugin for Datasette in a while. It pairs the MapLibre GL library with OpenFreeMap, the free vector map server run by Zsolt Ero, and combines what the datasette-cluster-map and datasette-leaflet-geojson plugins do: clustered latitude and longitude points plus GeoJSON shapes, drawn on [...]

## Exclusive: Wherobots AI agents automate complex spatial data engineering

DevFeed: [Exclusive: Wherobots AI agents automate complex spatial data engineering](<https://devfeed.tech/articles/exclusive-wherobots-ai-agents-automate-complex-spatial-data-engineering-140316.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/11/18/exclusive-wherobots-ai-agents-automate-complex-spatial-data-engineering/>)

Author: Paul Gillin

Published: 2026-11-18T11:00:42Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Data pipelines](<https://devfeed.tech/topics/data-pipelines.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [Geographic Information System](<https://devfeed.tech/topics/gis.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>), [Analytics Pipeline](<https://devfeed.tech/topics/analytics-pipeline.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [apache-sedona](<https://devfeed.tech/tags/apache-sedona.md>), [big-data](<https://devfeed.tech/tags/big-data.md>), [data-pipelines](<https://devfeed.tech/tags/data-pipelines.md>), [geographic-visualizations](<https://devfeed.tech/tags/geographic-visualizations.md>), [geospatial-data](<https://devfeed.tech/tags/geospatial-data.md>), [global-marketing-insights-inc](<https://devfeed.tech/tags/global-marketing-insights-inc.md>), [news](<https://devfeed.tech/tags/news.md>), [spatial](<https://devfeed.tech/tags/spatial.md>), [spatial-analytics](<https://devfeed.tech/tags/spatial-analytics.md>), [spatial-data-catalog](<https://devfeed.tech/tags/spatial-data-catalog.md>), [the-latest](<https://devfeed.tech/tags/the-latest.md>), [top-story-5](<https://devfeed.tech/tags/top-story-5.md>), [wherobots-havasu](<https://devfeed.tech/tags/wherobots-havasu.md>), [wherobots-inc](<https://devfeed.tech/tags/wherobots-inc.md>), [wherobots-rasterflow](<https://devfeed.tech/tags/wherobots-rasterflow.md>), [wherobots-waldo](<https://devfeed.tech/tags/wherobots-waldo.md>), [wherobotsdb](<https://devfeed.tech/tags/wherobotsdb.md>)

### AI overview

Wherobots' Waldo agent uses natural-language instructions to build spatial data pipelines, coordinating data discovery, query generation, analysis, visualization, and scheduled refreshes. Its platform combines Havasu for dataset discovery and preparation, RasterFlow for imagery processing, WherobotsDB with more than 300 spatial functions, and Wherobots Maps. Wherobots says it connects existing LLMs to specialized processing engines and can use existing cloud storage and catalogs such as AWS Glue and Databricks Unity Catalog, without requiring data migration. Users can inspect generated queries and processing steps, but the company cautions that prompts do not always produce accurate results. Waldo is in public preview; agent tokens are included at no extra charge in Professional Edition, while customers primarily pay for computing resources.

### Source excerpt

Cloud-native spatial analytics and artificial intelligence startup Wherobots Inc. today introduced an artificial intelligence agent that enables users to build and deploy spatial data pipelines using natural-language instructions, reducing the need for specialized programming expertise. Called Waldo, the agent can discover data sources, generate queries, perform analysis, visualize results and schedule recurring operations. It's designed [...] The post Exclusive: Wherobots AI agents automate complex spatial data engineering appeared first on SiliconANGLE.

## Ai2 and Taylor Geospatial partner to make geospatial AI easier to compare, run, and build on

DevFeed: [Ai2 and Taylor Geospatial partner to make geospatial AI easier to compare, run, and build on](<https://devfeed.tech/articles/ai2-and-taylor-geospatial-partner-to-make-geospatial-ai-easier-to-compare-run-and-build-on-141078.md>)

Original publisher: [Read original article](<https://allenai.org/blog/taylor-geospatial-partnership>)

Published: 2026-11-18T08:00:00Z

Content type: news

Language: en

Sources: [Ai2 Blog](<https://devfeed.tech/sources/ai2-blog.md>)

Topics: [geospatial](<https://devfeed.tech/topics/geospatial.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [AI Models](<https://devfeed.tech/topics/ai-models.md>), [Digital Public Good](<https://devfeed.tech/topics/digital-public-goods.md>), [test](<https://devfeed.tech/topics/test.md>), [human feedback](<https://devfeed.tech/topics/human-feedback.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [accelerate](<https://devfeed.tech/tags/accelerate.md>), [accessible](<https://devfeed.tech/tags/accessible.md>), [agriculture](<https://devfeed.tech/tags/agriculture.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai2](<https://devfeed.tech/tags/ai2.md>), [applications](<https://devfeed.tech/tags/applications.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [build](<https://devfeed.tech/tags/build.md>), [compare](<https://devfeed.tech/tags/compare.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [good](<https://devfeed.tech/tags/good.md>), [imagery](<https://devfeed.tech/tags/imagery.md>), [models](<https://devfeed.tech/tags/models.md>), [open](<https://devfeed.tech/tags/open.md>), [partner](<https://devfeed.tech/tags/partner.md>), [public](<https://devfeed.tech/tags/public.md>), [run](<https://devfeed.tech/tags/run.md>), [satellite](<https://devfeed.tech/tags/satellite.md>)

### AI overview

Ai2 and Taylor Geospatial are partnering to provide open resources that help practitioners use and evaluate geospatial AI for conservation, agriculture, and humanitarian work. The effort combines Benchmarks of The World for comparing models, Ai2's OlmoEarth open Earth-observation model family with a no-code platform to fine-tune and run models globally, and an expansion of Fields of The World to improve farm-field boundary precision and add crop type and yield. The partners say datasets, benchmarks, and model weights will be free under open licenses on Source Cooperative and GitHub. Fields of The World already uses a 13-member practitioner advisory group; the partners want similar input to shape shared benchmarks around field needs.

### Source excerpt

Ai2 and Taylor Geospatial are partnering to make geospatial AI more accessible and useful for the public good through open models, shared benchmarks, and improved satellite imagery datasets for applications in conservation, agriculture, and humanitarian work.

## DARWIN: a Geospatial Foundation Model for Marine Intelligence

DevFeed: [DARWIN: a Geospatial Foundation Model for Marine Intelligence](<https://devfeed.tech/articles/darwin-a-geospatial-foundation-model-for-marine-intelligence-140039.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22496>)

Author: Coen van den Elsen, Laurens Geffert

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [AI Development](<https://devfeed.tech/topics/ai-development.md>), [Earth AI](<https://devfeed.tech/topics/earth-ai.md>)

Tags: [2022](<https://devfeed.tech/tags/2022.md>), [anomaly-detection](<https://devfeed.tech/tags/anomaly-detection.md>), [autoencoder](<https://devfeed.tech/tags/autoencoder.md>), [challenges](<https://devfeed.tech/tags/challenges.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [data](<https://devfeed.tech/tags/data.md>), [embedding](<https://devfeed.tech/tags/embedding.md>), [environmental-monitoring](<https://devfeed.tech/tags/environmental-monitoring.md>), [error](<https://devfeed.tech/tags/error.md>), [foundation-model](<https://devfeed.tech/tags/foundation-model.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [global](<https://devfeed.tech/tags/global.md>), [marine](<https://devfeed.tech/tags/marine.md>), [masking](<https://devfeed.tech/tags/masking.md>), [model](<https://devfeed.tech/tags/model.md>), [modelling](<https://devfeed.tech/tags/modelling.md>), [physics-ao-ph](<https://devfeed.tech/tags/physics-ao-ph.md>), [random](<https://devfeed.tech/tags/random.md>), [spatial](<https://devfeed.tech/tags/spatial.md>), [structure](<https://devfeed.tech/tags/structure.md>), [temporal](<https://devfeed.tech/tags/temporal.md>)

### AI overview

DARWIN adapts a masked-autoencoder for marine data with missingness-aware encoding, expanded spatial and temporal masking, and Fourier-style positional embeddings. It is pretrained on 26 years of global Copernicus reanalysis and GEBCO bathymetry. For chlorophyll-a prediction, it reports the lowest monthly relative error in every month of 2022, outperforming no-encoder baselines and an identically architected model with random initialization. Trained with 10% of the supervised data, DARWIN also surpasses the random-init model trained with all of it. The abstract presents these results as evidence that the model learns coherent oceanographic structure and may serve as a backbone for downstream marine applications.

### Source excerpt

arXiv:2611.22496v1 Announce Type: new Abstract: We present DARWIN, a geospatial foundation model for marine intelligence. Recent Earth-observation foundation models have concentrated on terrestrial domains. The marine realm poses distinct challenges: missingness, multi-modal physical and biogeochemical variables, sparse labels, and dynamics that are transient, multi-scale, and non stationary over time. DARWIN adapts a masked-autoencoder backbone to these properties through missingness-aware encoding, expanded spatial and temporal masking, and Fourier-style positional embeddings, and is pretrained on 26 years of global Copernicus reanalysis together with GEBCO bathymetry. On predicting chlorophyll-a, the pretrained backbone attains the lowest monthly relative error in every month of 2022, outperforming both no-encoder baselines and an identically architected random-init encoder. At 10% of the supervised training data, DARWIN already surpasses the random-init model trained on 100% of the data. Embedding analysis and qualitative reconstructions show that DARWIN captures coherent oceanographic structure rather than memorising input statistics, providing a scalable backbone for downstream marine applications such as anomaly detection, species distribution modelling, and environmental monitoring.

## PRUE: A Practical Recipe for Field Boundary Segmentation at Scale

DevFeed: [PRUE: A Practical Recipe for Field Boundary Segmentation at Scale](<https://devfeed.tech/articles/prue-a-practical-recipe-for-field-boundary-segmentation-at-scale-141546.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2603.27101>)

Author: Gedeon Muhawenayo, Caleb Robinson, Subash Khanal, Zhanpei Fang, Isaac Corley, Alexander Wollam, Tianyi Gao, Leonard Strnad, Ryan Avery, Lyndon Estes, Ana M. T\\'arano, Nathan Jacobs, Hannah Kerner

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Image processing](<https://devfeed.tech/topics/image-processing.md>), [Advanced AI workflows](<https://devfeed.tech/topics/advanced-ai-workflows.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Dataset development](<https://devfeed.tech/topics/dataset-development.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [all](<https://devfeed.tech/tags/all.md>), [alternatives](<https://devfeed.tech/tags/alternatives.md>), [approach](<https://devfeed.tech/tags/approach.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [based](<https://devfeed.tech/tags/based.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [changes](<https://devfeed.tech/tags/changes.md>), [combines](<https://devfeed.tech/tags/combines.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [deep](<https://devfeed.tech/tags/deep.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [design](<https://devfeed.tech/tags/design.md>), [field](<https://devfeed.tech/tags/field.md>), [practical](<https://devfeed.tech/tags/practical.md>), [recipe](<https://devfeed.tech/tags/recipe.md>), [scale](<https://devfeed.tech/tags/scale.md>)

### AI overview

On the Fields of The World benchmark, a U-Net semantic segmentation model outperformed the evaluated instance-based and geospatial foundation models across performance and deployment metrics. The authors' approach combines a U-Net backbone with composite loss functions and targeted data augmentations to improve robustness to real-world variation. It reports 76% IoU and 47% object-F1, gains of 6 and 9 percentage points over the previous baseline, respectively. The abstract says the work releases all models and model-derived field-boundary datasets for five countries.

### Source excerpt

arXiv:2603.27101v3 Announce Type: replace-cross Abstract: Large-scale maps of field boundaries are essential for agricultural monitoring tasks. Existing deep learning approaches for satellite-based field mapping are sensitive to illumination, spatial scale, and changes in geographic location. We conduct the first systematic evaluation of segmentation and geospatial foundation models (GFMs) for global field boundary delineation using the Fields of The World (FTW) benchmark. We evaluate 18 models under unified experimental settings, showing that a U-Net semantic segmentation model outperforms instance-based and GFM alternatives on a suite of performance and deployment metrics. We propose a new segmentation approach that combines a U-Net backbone, composite loss functions, and targeted data augmentations to enhance performance and robustness under real-world conditions. Our model achieves a 76% IoU and 47% object-F1 on FTW, an increase of 6% and 9% over the previous baseline. Our approach provides a practical framework for reliable, scalable, and reproducible field boundary delineation across model design, training, and inference. We release all models and model-derived field boundary datasets for five countries.

## MAPS: A Multimodal Agentic Framework for POI Data Production Based on Crowdsourced Street-view Imagery

DevFeed: [MAPS: A Multimodal Agentic Framework for POI Data Production Based on Crowdsourced Street-view Imagery](<https://devfeed.tech/articles/maps-a-multimodal-agentic-framework-for-poi-data-production-based-on-crowdsourced-street-view-imagery-137496.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.19807>)

Author: Sijie Ruan, Tingyu Xie, Mingfei Liu, Yicheng Tang, Haoyu Han, Jingnan Qi, Jingshu Zeng, Dongqi Xu, Jing Wang, Shuliang Wang

Published: 2026-11-17T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>)

Topics: [smolvlm](<https://devfeed.tech/topics/smolvlm.md>), [Entity resolution](<https://devfeed.tech/topics/entity-resolution.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Image](<https://devfeed.tech/topics/image.md>), [managing large-scale text data in AI](<https://devfeed.tech/topics/managing-large-scale-text-data-in-ai.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [actions](<https://devfeed.tech/tags/actions.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [also](<https://devfeed.tech/tags/also.md>), [annotation](<https://devfeed.tech/tags/annotation.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [based](<https://devfeed.tech/tags/based.md>), [but](<https://devfeed.tech/tags/but.md>), [candidate](<https://devfeed.tech/tags/candidate.md>), [changes](<https://devfeed.tech/tags/changes.md>), [completeness](<https://devfeed.tech/tags/completeness.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [data](<https://devfeed.tech/tags/data.md>), [database](<https://devfeed.tech/tags/database.md>), [databases](<https://devfeed.tech/tags/databases.md>), [framework](<https://devfeed.tech/tags/framework.md>), [imagery](<https://devfeed.tech/tags/imagery.md>), [maps](<https://devfeed.tech/tags/maps.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [production](<https://devfeed.tech/tags/production.md>), [view](<https://devfeed.tech/tags/view.md>)

### AI overview

MAPS turns crowdsourced street-view imagery into POI database updates through two cascaded modules. SPICE adapts a lightweight vision-language model to reconstruct and filter text, group entries by POI, and classify their meaning and completeness. RIMA retrieves candidate records, matches POIs using evidence and reasons about attributes, selectively gathering additional visual, geospatial, or external-source evidence when needed. In a human-reviewed full-task evaluation, MAPS outperformed the strongest baseline by 21.8%; deployment in Kuaishou's large-scale POI pipeline reduced average processing time by 76.9% compared with manual annotation.

### Source excerpt

arXiv:2611.19807v1 Announce Type: new Abstract: Point-of-interest (POI) databases support location-based services, yet urban changes quickly make their records stale and manual maintenance is costly. Crowdsourced street-view images provide timely visual evidence, but converting them into database actions requires separating multiple POIs and irrelevant text in each image, then resolving incomplete observations against ambiguous records. We study Street-view Image-based POI Data Production and propose MAPS, the first framework to integrate structured POI extraction with database maintenance from such imagery. MAPS contains two cascaded modules. SPICE progressively adapts a lightweight vision-language model to reconstruct and filter text, group entries belonging to the same POI, and classify their semantics and completeness. RIMA retrieves candidate records, performs evidence-grounded POI matching and attribute-level reasoning, and selectively acquires complementary visual, geospatial, or external-source evidence when necessary. In human-reviewed full-task evaluation, MAPS outperforms the strongest baseline by 21.8\%. MAPS has also been deployed in Kuaishou's large-scale POI production pipeline, reducing average processing time by 76.9\% relative to manual annotation. These results demonstrate the effectiveness and practical value of MAPS for scalable POI maintenance.

## DIAGNO: Diagonal Spherical Neural Operators for Heterogeneous Earth Dynamics Modeling

DevFeed: [DIAGNO: Diagonal Spherical Neural Operators for Heterogeneous Earth Dynamics Modeling](<https://devfeed.tech/articles/diagno-diagonal-spherical-neural-operators-for-heterogeneous-earth-dynamics-modeling-132225.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.16356>)

Author: Herui Li, Bin Lu, Haonan Qi, Lei Zhou, Luoyi Fu, Xinbing Wang, Meng Jin

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>)

Tags: [atmosphere](<https://devfeed.tech/tags/atmosphere.md>), [capture](<https://devfeed.tech/tags/capture.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data-driven](<https://devfeed.tech/tags/data-driven.md>), [earth](<https://devfeed.tech/tags/earth.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fluid-dynamics](<https://devfeed.tech/tags/fluid-dynamics.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [interaction](<https://devfeed.tech/tags/interaction.md>), [kernel](<https://devfeed.tech/tags/kernel.md>), [modeling](<https://devfeed.tech/tags/modeling.md>), [neural](<https://devfeed.tech/tags/neural.md>), [ocean](<https://devfeed.tech/tags/ocean.md>), [operators](<https://devfeed.tech/tags/operators.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [rotation](<https://devfeed.tech/tags/rotation.md>), [solve](<https://devfeed.tech/tags/solve.md>), [spectrum](<https://devfeed.tech/tags/spectrum.md>), [state](<https://devfeed.tech/tags/state.md>), [system](<https://devfeed.tech/tags/system.md>)

### AI overview

DIAGNO targets a limitation of spherical neural operators: rotation equivariance isolates spectral modes and can over-idealize Earth dynamics as homogeneous. It derives an explicit modal interaction kernel and parameterizes cross-modal interactions, with inter- and intra-diagonal mechanisms intended to capture zonal and meridional dynamics. The abstract reports state-of-the-art multi-step forecasting across simulated fluid dynamics and atmosphere and ocean reanalysis datasets, and says evaluations found fidelity in climatological anomalies and physical consistency. These are claims from the abstract; it provides no specific metrics or implementation details.

### Source excerpt

arXiv:2611.16356v1 Announce Type: new Abstract: Accurate forecasting of Earth's dynamic systems is vital in geophysics, where neural operators offer a powerful data-driven paradigm. Standard operators often treat the Earth as a flat Euclidean plane, introducing geometric artifacts. In contrast, Spherical Neural Operators (SNOs) resolve this mismatch by operating within the spherical spectral space, naturally adapting to spherical geometries. However, for computational efficiency, existing SNOs typically rely on the rotation equivariance assumption. This assumption simplifies the transitions of physical states into isolated spectral modal evolutions, thereby over-idealizing the Earth as a homogeneous system. To solve this problem, we no longer rely on the rotation equivariance assumption that isolates spectral modal evolution, but directly derive an explicit modal interaction kernel and propose Diagonal Spherical Neural Operators named DIAGNO. DIAGNO explicitly parameterizes cross-modal interactions through spectral modal decomposition, thereby preserving the inherent heterogeneity of Earth dynamics. Furthermore, leveraging the unique diagonal structure of the spherical spectrum, we design inter- and intra-diagonal interaction mechanisms to natively capture essential zonal and meridional dynamics, respecting intrinsic geophysical regularities. Extensive experiments across simulated fluid dynamics and real-world Earth system reanalysis (including atmosphere and ocean) datasets demonstrate DIAGNO's superior capability, consistently achieving state-of-the-art performance in multi-step forecasting. Beyond this, we validate DIAGNO's practical analytical value in professional geophysical domains by assessing its fidelity in capturing climatological anomalies and maintaining physical consistency.

## STRATOS: Bridging the Symbolic-to-Numeric Gap in Spatio-Temporal Text-to-SQL for Meteorological Data

DevFeed: [STRATOS: Bridging the Symbolic-to-Numeric Gap in Spatio-Temporal Text-to-SQL for Meteorological Data](<https://devfeed.tech/articles/stratos-bridging-the-symbolic-to-numeric-gap-in-spatio-temporal-text-to-sql-for-meteorological-data-131944.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2607.03501>)

Author: Yi Zhang, Farhad Nooralahzadeh, Jonathan F\\"urst, Fabio Scherrer, Claude Lehmann, Antonis Bezes, Vassiliki Kotroni, Kurt Stockinger

Published: 2026-11-13T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>)

Topics: [text2sql](<https://devfeed.tech/topics/text2sql.md>), [AI for SQL query generation](<https://devfeed.tech/topics/ai-for-sql-query-generation.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [SQL](<https://devfeed.tech/topics/sql.md>), [Dataset development](<https://devfeed.tech/topics/dataset-development.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [climate](<https://devfeed.tech/tags/climate.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [cs-ma](<https://devfeed.tech/tags/cs-ma.md>), [data](<https://devfeed.tech/tags/data.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [dimensions](<https://devfeed.tech/tags/dimensions.md>), [execution](<https://devfeed.tech/tags/execution.md>), [exploration](<https://devfeed.tech/tags/exploration.md>), [external](<https://devfeed.tech/tags/external.md>), [framework](<https://devfeed.tech/tags/framework.md>), [knowledge](<https://devfeed.tech/tags/knowledge.md>), [mapping](<https://devfeed.tech/tags/mapping.md>), [ontology](<https://devfeed.tech/tags/ontology.md>), [query](<https://devfeed.tech/tags/query.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [scalability](<https://devfeed.tech/tags/scalability.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [spatial](<https://devfeed.tech/tags/spatial.md>), [sql](<https://devfeed.tech/tags/sql.md>), [temporal](<https://devfeed.tech/tags/temporal.md>), [text](<https://devfeed.tech/tags/text.md>)

### AI overview

STRATOS aims to make meteorological datasets more accessible through natural-language Text-to-SQL by resolving spatial and semantic ambiguities before SQL generation. Its resolution agent maps fuzzy concepts to a localized ontology and uses external knowledge bases to resolve spatial entities; a complexity-aware query rewriter changes expensive spatial predicates, with the abstract reporting execution times reduced from hours to seconds. The authors also introduce a domain-expert workload of 7,520 complex query pairs to evaluate scalability and symbolic-to-numeric translation across spatio-temporal dimensions. These are claims in the supplied abstract; it does not provide evaluation details or further evidence for the reported speedup.

### Source excerpt

arXiv:2607.03501v2 Announce Type: replace-cross Abstract: Copernicus, the European Union's Earth observation program, produces petabytes of Earth observation and climate data, offering immense potential for research, policy, and applications. However, access to these datasets requires advanced programming skills and familiarity with domain-specific formats such as NetCDF or GRIB. Moreover, general-purpose Text-to-SQL systems fail when applied naively to the meteorological domain due to a profound ``Symbolic-to-Numeric'' gap. To overcome these limitations, we present an end-to-end Text-to-SQL framework specifically engineered for real-world, scalable meteorological data exploration. Our system intercepts natural language to resolve spatial and semantic ambiguities \textit{before} SQL generation. We design STRATOS, a Spatio-Temporal Resolution Agent for Text-to-SQL to dynamically bridge the symbolic-to-numeric gap by mapping fuzzy concepts to a localized ontology and resolving spatial entities via external knowledge bases. Further, our complexity-aware query rewriter rewrites expensive spatial predicates, reducing execution times from hours to seconds. Last, we introduce the STRATOS Evaluation Workload, comprising 7,520 complex query pairs explicitly designed by domain experts to test scalability and symbolic-to-numeric translation across challenging spatio-temporal dimensions previously unexplored by Text-to-SQL systems.

## Satellite Pretraining Is Not One Thing: A Matched-Pair Study of Generic and Earth-Observation Foundation Models for Scene Classification

DevFeed: [Satellite Pretraining Is Not One Thing: A Matched-Pair Study of Generic and Earth-Observation Foundation Models for Scene Classification](<https://devfeed.tech/articles/satellite-pretraining-is-not-one-thing-a-matched-pair-study-of-generic-and-earth-observation-foundation-models-for-scene-classification-128363.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.13849>)

Author: Sandarbh Agrawal, Rakshith Sathish, Ashwathi Nambiar

Published: 2026-11-11T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [geospatial](<https://devfeed.tech/topics/geospatial.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [image](<https://devfeed.tech/tags/image.md>), [pretraining](<https://devfeed.tech/tags/pretraining.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [rgb](<https://devfeed.tech/tags/rgb.md>), [satellite](<https://devfeed.tech/tags/satellite.md>), [small](<https://devfeed.tech/tags/small.md>), [target](<https://devfeed.tech/tags/target.md>)

### AI overview

Across five matched pairs of generic and Earth-observation encoders, satellite-image pretraining did not reliably improve few-label scene classification. On most datasets and metrics, SatDINO, Scale-MAE and RemoteCLIP improved results, while DINOv3-SAT and SatMAE hurt linear-probe accuracy by up to 22.5 and 8.8 points, respectively; even two MAE models pretrained on the same fMoW-RGB imagery had opposite outcomes. Effects were strongest with 1-5% labels, concentrated on small man-made objects, and persisted with patch-token pooling and attentive probes. Frozen-embedding silhouette scores closely tracked k-NN accuracy and retrieval mAP on four of five datasets. The authors conclude that checkpoints need validation on target data; satellite pretraining alone is not a dependable selection criterion.

### Source excerpt

arXiv:2611.13849v1 Announce Type: cross Abstract: Geospatial foundation models are pretrained on satellite and aerial imagery, but whether such an encoder beats a strong generic one for few-label land-use/land-cover mapping remains unclear, as prior comparisons rarely isolate the pretraining data. We study five matched pairs of Earth-observation (EO) and generic encoders sharing objective and architecture (DINO/SatDINO, DINOv3/DINOv3-SAT, MAE/SatMAE, MAE/Scale-MAE, CLIP/RemoteCLIP) within a benchmark of 14 frozen encoders on five RGB scene-classification datasets (0.06-30 m GSD). With native preprocessing, we evaluate k-NN, linear and attentive probes and image retrieval over six low-label splits, using paired per-image significance tests. EO pretraining is not uniformly helpful: on most datasets and metrics, it significantly improves three pairs (SatDINO, Scale-MAE, RemoteCLIP, the latter evaluated only on unseen datasets) and significantly hurts two (linear probe: DINOv3-SAT by up to 22.5 points, SatMAE by up to 8.8), with opposite outcomes even for two MAEs pretrained on the same fMoW-RGB imagery. These effects are robust to patch-token pooling and attentive probing, concentrate on small man-made objects, and are largest with 1-5% labels. Across encoders, the silhouette score of frozen embeddings predicts k-NN accuracy and retrieval mAP almost perfectly (Spearman $\rho \geq 0.96$ on four of five datasets). Pretraining on satellite imagery is thus not a reliable selection criterion; candidate checkpoints must be validated on target data.

## Multimodal Road Closure Mapping via Cross-Modal Graph Attention

DevFeed: [Multimodal Road Closure Mapping via Cross-Modal Graph Attention](<https://devfeed.tech/articles/multimodal-road-closure-mapping-via-cross-modal-graph-attention-128266.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.13192>)

Author: Izzeddin Teeti, Vaibhav Kango, Nidhi Sridhar, Mohamed Moustafa

Published: 2026-11-11T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [multimodal-ai](<https://devfeed.tech/topics/multimodal-ai.md>), [Transformer](<https://devfeed.tech/topics/transformer.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Neural Network](<https://devfeed.tech/topics/neural-network.md>)

Tags: [association](<https://devfeed.tech/tags/association.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [classification](<https://devfeed.tech/tags/classification.md>), [closure](<https://devfeed.tech/tags/closure.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [framework](<https://devfeed.tech/tags/framework.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [graph](<https://devfeed.tech/tags/graph.md>), [imagery](<https://devfeed.tech/tags/imagery.md>), [inductive](<https://devfeed.tech/tags/inductive.md>), [inference](<https://devfeed.tech/tags/inference.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [map](<https://devfeed.tech/tags/map.md>), [mapping](<https://devfeed.tech/tags/mapping.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [network](<https://devfeed.tech/tags/network.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [satellite](<https://devfeed.tech/tags/satellite.md>), [spatial](<https://devfeed.tech/tags/spatial.md>), [structured](<https://devfeed.tech/tags/structured.md>)

### AI overview

The framework jointly classifies road closures as none, partial, or complete and associates them with road segments. It combines graph structured map data, satellite imagery augmented with projected street level detections, and street view images using modality specific encoders and cross modal transformer attention. On nearly 20,000 real world fleet scenarios, it reports 72.7% closure classification accuracy and 60% segment association accuracy. The abstract says it exceeds frontier vision language models by over 28 percentage points in closure classification while being over 4,000 times smaller and 100 times faster at inference. Ablations identify cross modal attention, edge aware pooling, and graph message passing as important to performance.

### Source excerpt

arXiv:2611.13192v1 Announce Type: cross Abstract: Accurate and timely identification of road closures is critical for large-scale navigation and logistics systems, yet remains challenging due to the need to jointly reason over visual evidence, spatial context, and road network topology. Existing approaches either rely on manual map updates or focus on detecting closure-related objects without resolving which road segment is affected. In this work, we introduce a unified multimodal and multitask framework that jointly performs road closure classification (no, partial, complete) and road segment association within a single learning architecture. Our model integrates graph-structured map data, satellite imagery augmented with projected street-level detections, and street-view images using modality-specific encoders and cross-modal transformer attention, enabling explicit reasoning over visual cues and network topology. Evaluated on nearly 20K real-world scenarios collected from a large-scale commercial vehicle fleet, our approach achieves 72.7% closure classification accuracy and 60% road segment association accuracy, outperforming frontier vision-language models by over 28 percentage points on closure classification while being over 4,000x smaller and 100x faster at inference. Extensive ablations demonstrate the importance of cross-modal attention, edge-aware pooling, and graph message passing for accurate geospatial reasoning. This work establishes the first benchmark for joint road closure classification and segment-level attribution, and shows that task-specific architectural inductive biases substantially outperform general-purpose multimodal models for structured geospatial understanding.

## Opening the datasets behind our AI weather models

DevFeed: [Opening the datasets behind our AI weather models](<https://devfeed.tech/articles/opening-the-datasets-behind-our-ai-weather-models-133528.md>)

Original publisher: [Read original article](<https://deepmind.google/blog/opening-the-datasets-behind-our-ai-weather-models/>)

Author: WeatherNext team

Published: 2026-11-10T11:13:04Z

Content type: release

Language: en

Sources: [Google DeepMind News](<https://devfeed.tech/sources/google-deepmind-news.md>)

Topics: [Dataset development](<https://devfeed.tech/topics/dataset-development.md>), [Climate / Weather / Ocean Modeling](<https://devfeed.tech/topics/climate-weather-ocean-modeling.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [applications](<https://devfeed.tech/tags/applications.md>), [climate](<https://devfeed.tech/tags/climate.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [developers](<https://devfeed.tech/tags/developers.md>), [earth](<https://devfeed.tech/tags/earth.md>), [global](<https://devfeed.tech/tags/global.md>), [imagery](<https://devfeed.tech/tags/imagery.md>), [live](<https://devfeed.tech/tags/live.md>), [modeling](<https://devfeed.tech/tags/modeling.md>), [models](<https://devfeed.tech/tags/models.md>), [satellite](<https://devfeed.tech/tags/satellite.md>), [satellites](<https://devfeed.tech/tags/satellites.md>), [science](<https://devfeed.tech/tags/science.md>), [train](<https://devfeed.tech/tags/train.md>), [water](<https://devfeed.tech/tags/water.md>), [weather](<https://devfeed.tech/tags/weather.md>)

### AI overview

SatWeave and PARDIG are being released with historical archives and live feeds for researchers and developers building weather, hydrology, and climate applications. SatWeave combines imagery from five geostationary satellites into a global mosaic that gives WeatherNext 3 a live view of clouds and storms. The satellites differ in map projections, file formats, sensors, and scan schedules of 10 to 30 minutes, making a coherent global map difficult; the excerpt says existing composites commonly omit incompatible details, use four or fewer channels, or update only every few hours. PARDIG is an AI-based global precipitation dataset used to train WeatherNext 3 precipitation forecasts.

### Source excerpt

Foundational precipitation and geostationary satellite datasets to observe the planet clearly.

## Decomposing Earth Embeddings with Sparse Autoencoders

DevFeed: [Decomposing Earth Embeddings with Sparse Autoencoders](<https://devfeed.tech/articles/decomposing-earth-embeddings-with-sparse-autoencoders-125375.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.11398>)

Author: Vitus Benson, Fanny Yang, Markus Reichstein

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>)

Tags: [agriculture](<https://devfeed.tech/tags/agriculture.md>), [autoencoders](<https://devfeed.tech/tags/autoencoders.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [earth](<https://devfeed.tech/tags/earth.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [features](<https://devfeed.tech/tags/features.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [physics-geo-ph](<https://devfeed.tech/tags/physics-geo-ph.md>), [robustness](<https://devfeed.tech/tags/robustness.md>), [sparsity](<https://devfeed.tech/tags/sparsity.md>), [spatial](<https://devfeed.tech/tags/spatial.md>)

### AI overview

Sparse autoencoders (SAEs) make Earth observation embeddings more interpretable by decomposing them into sparse features, and they reconstruct AlphaEarth and Tessera embeddings better than the tested baselines at matched sparsity. The authors evaluate reconstruction, alignment with reference concepts, and spatial scale, including comparison with an SAE using a frozen random decoder. On AlphaEarth, 88% of learned features align with at most one of 20 concept groups, while raw dimensions have a median of 12; BatchTopK and Matryoshka features also place 23-32% of variance within a 640 m patch, versus 9% for raw embeddings. The random-decoder SAE shares the concept-alignment and fine-scale properties, suggesting those arise from sparsity, but learned dictionaries reconstruct more accurately: trained SAEs leave eight times less variance unexplained on AlphaEarth. Tests of land-cover sub-concept discovery and out-of-distribution detection show potential, though neither application is yet reliable enough for practical use.

### Source excerpt

arXiv:2611.11398v1 Announce Type: new Abstract: Earth observation foundation models produce global metre-resolution embeddings that underpin work across agriculture, ecology, hydrology, and beyond. Yet these embeddings are opaque: each dimension entangles unrelated physical concepts, which makes them difficult to interpret and their robustness hard to assess. Here, we propose sparse autoencoders (SAEs) as a principled way to decompose Earth embeddings into sparse, approximately monosemantic features over an overcomplete dictionary. Because LLM judges of interpretations do not transfer to satellite pixels, and spatial autocorrelation matters, we introduce a tailored evaluation protocol. It scores any SAE on reconstruction at matched sparsity, alignment with reference concepts, and spatial scale, each against baselines that include an SAE whose decoder is a frozen random dictionary. On this protocol, SAEs reconstruct AlphaEarth and Tessera embeddings better than every baseline at matched sparsity. Their features align with far fewer reference concepts than raw embedding dimensions (on AlphaEarth, 88% align with at most one of 20 concept groups, most with none, against a median of 12 per raw dimension). They also carry more of their variance at fine spatial scales (23 to 32% within a 640 m patch for BatchTopK and Matryoshka, against 9%). An SAE with a frozen random decoder shares both properties, which therefore follow from sparsity, whereas only a learned dictionary reconstructs the embedding accurately: on AlphaEarth, trained SAEs leave eight times less variance unexplained. We test the sparse code on two applications, discovering land-cover sub-concepts and detecting out-of-distribution pixels by area-of-applicability estimation. Both indicate how sparse features could be used, but neither is yet reliable enough for practical use. Together, these results are a step toward Earth observation models whose features can be inspected.

## Global Solar Siting from Geospatial Embeddings

DevFeed: [Global Solar Siting from Geospatial Embeddings](<https://devfeed.tech/articles/global-solar-siting-from-geospatial-embeddings-123311.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.08831>)

Author: Abdullah Rehman, Quentin Paletta, Rui Song

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [mlp](<https://devfeed.tech/topics/mlp.md>)

Tags: [capacity](<https://devfeed.tech/tags/capacity.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-cy](<https://devfeed.tech/tags/cs-cy.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [deployment](<https://devfeed.tech/tags/deployment.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [energy](<https://devfeed.tech/tags/energy.md>), [existing](<https://devfeed.tech/tags/existing.md>), [future](<https://devfeed.tech/tags/future.md>), [generation](<https://devfeed.tech/tags/generation.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [global](<https://devfeed.tech/tags/global.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [installations](<https://devfeed.tech/tags/installations.md>), [models](<https://devfeed.tech/tags/models.md>), [perceptron](<https://devfeed.tech/tags/perceptron.md>), [practical](<https://devfeed.tech/tags/practical.md>), [properties](<https://devfeed.tech/tags/properties.md>), [pv](<https://devfeed.tech/tags/pv.md>), [random](<https://devfeed.tech/tags/random.md>), [ranking](<https://devfeed.tech/tags/ranking.md>), [solar](<https://devfeed.tech/tags/solar.md>)

### AI overview

The study uses geospatial embeddings of 38,591 existing solar sites to train a soft-voting random forest and multilayer perceptron ensemble that ranks land by solar suitability. Tested against later installations in 13 countries, it achieved a median ROC-AUC of 0.90 and exceeded published multi-criteria decision analysis strategies by 0.06-0.30 ROC-AUC in every country. The top-ranked 20% of land contains 84% of estimated generation and 67% of installed capacity. After standard siting constraints, 5.6% of global land remains highly suited, rising to 9.8% when cropland is allowed. The abstract says a global 10 m suitability dataset and code are available.

### Source excerpt

arXiv:2611.08831v1 Announce Type: new Abstract: Global solar photovoltaic (PV) capacity is rapidly growing, helping decrease our emissions and dependence on fossil fuels. However, optimally siting a solar plant remains challenging, requiring anticipation of geographical, urban, and infrastructure needs. Current methods rely on multi-criteria decision analysis (MCDA) using hand-picked geophysical layers and solar-resource products, however these strategies are not geographically transferable and do not capture the practical requirements of solar sites. We use geospatial embeddings, dense vectors representing geophysical properties of a location, to analyse 38,591 existing solar sites and the land they were built on to learn the latent solar suitability signal. We assess a set of linear and non-linear models and select a soft-voting ensemble of a random forest and a multilayer perceptron (RF+MLP ensemble), chosen for accuracy and large scale deployability. Validated against future installations in thirteen countries, the ensemble separates existing solar sites from random land with a median ROC-AUC of 0.90. Compared with published MCDA strategies, the ensemble increases ROC-AUC by 0.06-0.30 in every country. The ranking also concentrates real deployment: the most-suitable 20% of land holds 84% of estimated generation (annual energy yield) and 67% of installed capacity (potential power output). After applying standard siting constraints, 5.6% of global land (9.8% if cropland is permitted) remains highly suited to solar. The global 10 m resolution solar suitability dataset and code can be found here: https://github.com/Abdullahr29/Global-Solar-Siting.

## Towards Label-Efficient Geospatial Foundation Model Adaptation via Positive-Unlabeled Learning for Building Damage Mapping

DevFeed: [Towards Label-Efficient Geospatial Foundation Model Adaptation via Positive-Unlabeled Learning for Building Damage Mapping](<https://devfeed.tech/articles/towards-label-efficient-geospatial-foundation-model-adaptation-via-positive-unlabeled-learning-for-building-damage-mapping-124530.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.08561>)

Author: Hengwei Zhao, Qirui Wang, Zhuo Zheng, Xiaoyan Lu, Rusty Feagin, Wenzhe Jiao

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [geospatial](<https://devfeed.tech/topics/geospatial.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [peft](<https://devfeed.tech/topics/peft.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Representation learning](<https://devfeed.tech/topics/representation-learning.md>)

Tags: [2025](<https://devfeed.tech/tags/2025.md>), [annotation](<https://devfeed.tech/tags/annotation.md>), [annotations](<https://devfeed.tech/tags/annotations.md>), [application](<https://devfeed.tech/tags/application.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [building](<https://devfeed.tech/tags/building.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [data](<https://devfeed.tech/tags/data.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [disaster](<https://devfeed.tech/tags/disaster.md>), [earth](<https://devfeed.tech/tags/earth.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [foundation](<https://devfeed.tech/tags/foundation.md>), [foundation-model](<https://devfeed.tech/tags/foundation-model.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [learning](<https://devfeed.tech/tags/learning.md>), [mapping](<https://devfeed.tech/tags/mapping.md>), [model](<https://devfeed.tech/tags/model.md>)

### AI overview

PUMA adapts geospatial foundation models for building-damage mapping using annotations for only a small number of damaged buildings. It combines non-contrastive weakly supervised representation refinement, intended to reduce noisy supervision's effect on learning, with latent-state label disambiguation that refines supervision reversibly without estimating class priors. The abstract reports effective, robust results on disaster benchmarks and a 2025 Los Angeles wildfire application under highly limited supervision. It identifies discriminative representation learning as a challenge when using weak supervision; the supplied evidence does not provide quantitative results or implementation details.

### Source excerpt

arXiv:2611.08561v1 Announce Type: cross Abstract: Geospatial foundation models (GFMs) are reshaping remote sensing by shifting the field from task-specific modeling toward general-purpose geospatial intelligence. Mapping the damage to buildings stands to benefit substantially from this shift, as GFMs pretrained on large-scale Earth observation data provide generalizable geospatial representations that can facilitate effective responses to sudden-onset disasters. Although parameter-efficient fine-tuning (PEFT) facilitates GFM adaptation, parameter efficiency does not necessarily translate into label efficiency. Building damage mapping remains challenging due to scarce pre-existing labeled data or limited time for annotation. When weak supervision is introduced to alleviate the labeling bottleneck, we identify the difficulty of learning sufficiently discriminative representations as a key challenge for label-efficient GFM adaptation. To address this challenge, we propose a Positive-Unlabeled GFM Adaptation (PUMA) framework that requires annotations for only a small number of damaged buildings. More specifically, PUMA integrates non-contrastive weakly supervised representation refinement to mitigate the impact of noisy supervision on representation learning from a gradient perspective, and latent-state-based label disambiguation to enable reversible supervision refinement without requiring class prior estimation. Extensive experiments on disaster benchmark datasets from around the world demonstrated the effectiveness and robustness of PUMA. A real-world application to the 2025 Los Angeles wildfires further highlighted its potential to accurately map damage under highly limited supervision. More broadly, these findings suggest that weakly supervised PEFT offers a label-efficient pathway from GFMs to real-world applications.

## Utilizing Earth Foundation Models to Enhance the Simulation Performance of Hydrological Models with AlphaEarth Embeddings

DevFeed: [Utilizing Earth Foundation Models to Enhance the Simulation Performance of Hydrological Models with AlphaEarth Embeddings](<https://devfeed.tech/articles/utilizing-earth-foundation-models-to-enhance-the-simulation-performance-of-hydrological-models-with-alphaearth-embeddings-118186.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2601.01558>)

Author: Pengfei Qu, Lei Ye, Wenyu Ouyang, Chi Zhang, Yikai Chai, Shuolong Xu, Yongri Piao, Miao Zhang, Huchuan Lu

Published: 2026-11-05T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Simulation](<https://devfeed.tech/topics/simulation.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [advanced](<https://devfeed.tech/tags/advanced.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [based](<https://devfeed.tech/tags/based.md>), [behavior](<https://devfeed.tech/tags/behavior.md>), [but](<https://devfeed.tech/tags/but.md>), [cluster](<https://devfeed.tech/tags/cluster.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [deep](<https://devfeed.tech/tags/deep.md>), [earth](<https://devfeed.tech/tags/earth.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [foundation](<https://devfeed.tech/tags/foundation.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [models](<https://devfeed.tech/tags/models.md>), [performance](<https://devfeed.tech/tags/performance.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

Using the CAMELS dataset, the study evaluates 64-dimensional AlphaEarth Foundations (AEF) embeddings as static basin descriptors for streamflow prediction. AEF embeddings had nearly the same in-sample performance as conventional CAMELS attributes but greater out-of-sample robustness, and generally improved prediction across 100 target basins in ungauged-basin experiments. With AEF inputs, similarity-based donor selection beat random selection for smaller donor pools; that advantage faded as pools grew, and adding donors did not consistently help. Leave-one-cluster-out results showed that benefits varied with basin partitioning, so transferability remained sensitive to which catchment conditions were represented during training. The embeddings were statistically associated with conventional terrain, vegetation, and rooting attributes.

### Source excerpt

arXiv:2601.01558v3 Announce Type: replace-cross Abstract: Accurate prediction in ungauged basins (PUB) requires effective transfer across hydrologically diverse regions. While deep learning has advanced large-sample hydrological modeling, its cross-regional generalization depends in part on the static basin descriptors used to guide model behavior. Earth Foundation Models (EFMs) provide a promising alternative by learning environmental representations from multi-modal satellite data. In this study, we evaluate the 64-dimensional AlphaEarth Foundations (AEF) embeddings as a new form of static basin representation and assess their value for streamflow prediction under both in-sample (IS) and out-of-sample (OOS) settings using the CAMELS dataset. Results show that AEF embeddings offer higher OOS robustness than CAMELS attributes, despite nearly identical IS performance. Mutual-information analysis reveals statistical associations between AEF embeddings and conventional catchment attributes related to terrain, vegetation, and rooting characteristics. We further investigate similarity-based donor-basin selection for PUB using CAMELS attributes and AEF embeddings. Across 100 target basins, AEF static inputs generally improved PUB performance. With AEF inputs, similarity-based donor selection outperformed random selection for smaller donor pools, but its advantage diminished as pool size increased. Adding more donor basins did not consistently improve performance. Leave-one-cluster-out experiments showed AEF input benefits depended on basin partitioning, suggesting that transferability remains sensitive to the catchment conditions represented during training. Overall, this study demonstrates that Earth Foundation Model embeddings provide a transferable representation for PUB prediction and underscores the importance of basin representation quality and donor pool design, motivating future research toward unified hydrological foundation models.

## COL-Trees: Efficient Hierarchical Object Search in Spatial Networks

DevFeed: [COL-Trees: Efficient Hierarchical Object Search in Spatial Networks](<https://devfeed.tech/articles/col-trees-efficient-hierarchical-object-search-in-spatial-networks-118104.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2601.22183>)

Author: Tenindra Abeywickrama, Muhammad Aamir Cheema, Sabine Storandt

Published: 2026-11-05T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>)

Topics: [graph theory](<https://devfeed.tech/topics/graph-theory.md>), [ann](<https://devfeed.tech/topics/ann.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>)

Tags: [algorithms](<https://devfeed.tech/tags/algorithms.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [cs-ds](<https://devfeed.tech/tags/cs-ds.md>), [data-structure](<https://devfeed.tech/tags/data-structure.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [edge](<https://devfeed.tech/tags/edge.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [graph](<https://devfeed.tech/tags/graph.md>), [improvement](<https://devfeed.tech/tags/improvement.md>), [magnitude](<https://devfeed.tech/tags/magnitude.md>), [multiple-agents](<https://devfeed.tech/tags/multiple-agents.md>), [nearest-neighbor](<https://devfeed.tech/tags/nearest-neighbor.md>), [networks](<https://devfeed.tech/tags/networks.md>), [overhead](<https://devfeed.tech/tags/overhead.md>), [queries](<https://devfeed.tech/tags/queries.md>), [query](<https://devfeed.tech/tags/query.md>), [real-world](<https://devfeed.tech/tags/real-world.md>), [search](<https://devfeed.tech/tags/search.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>), [theory](<https://devfeed.tech/tags/theory.md>), [trees](<https://devfeed.tech/tags/trees.md>)

### AI overview

COL-Tree (Compacted Object-Landmark Tree) supports hierarchical graph traversal with a landmark-based heuristic for spatial-network searches. The authors use it to answer Aggregate k Nearest Neighbor (A$k$NN), k Farthest Neighbors (k$FN), and other queries, which single-agent nearest-neighbor heuristics and Euclidean-based approaches may handle poorly in networks. Experiments on real and synthetic road, terrain, and 2D physics mesh networks report up to three orders of magnitude improvement over existing approaches, with small preprocessing overhead. These claims are from the supplied abstract.

### Source excerpt

arXiv:2601.22183v2 Announce Type: replace Abstract: Many spatial network applications, such as map-based services, rely heavily on efficient methods to search for relevant nearby points-of-interest (POIs). A $k$ Nearest Neighbor ($k$NN) query is one such example that finds the $k$ closest POIs from an agent's location. While most existing techniques for spatial networks focus on retrieving nearby POIs for a single agent, these search heuristics do not translate to many other applications. For example, Aggregate $k$ Nearest Neighbor (A$k$NN) queries require POIs that are close to multiple agents. $k$ Farthest Neighbors ($k$FN) queries require POIs that are the antithesis of nearest. Such problems naturally benefit from a hierarchical approach, but existing methods rely on Euclidean-based heuristics, which have diminished effectiveness in spatial networks such as road network graphs. We propose a novel data structure, COL-Tree (Compacted Object-Landmark Tree), to address this gap by enabling efficient hierarchical graph traversal using a more accurate landmark-based heuristic. We then present query algorithms that utilize COL-Trees to efficiently answer A$k$NN, $k$FN, and other queries. Our experiments evaluate performance on real-world and synthetic datasets including multiple types of spatial networks, such as road networks of varying sizes and edge weights, terrain networks, and 2D physics meshes. The results demonstrate that our techniques significantly outperform existing approaches, achieving up to three orders of magnitude improvement. Moreover, this comes at a small pre-processing overhead in both theory and practice.

## Generative AI for urban planning: synthesizing land-use maps with satellite imagery via DualControlNet

DevFeed: [Generative AI for urban planning: synthesizing land-use maps with satellite imagery via DualControlNet](<https://devfeed.tech/articles/generative-ai-for-urban-planning-synthesizing-land-use-maps-with-satellite-imagery-via-dualcontrolnet-118392.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.05859>)

Author: Qingqi Song, Yuzhou Chen, Lingqian Hu, Shenhao Wang

Published: 2026-11-05T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Generative AI](<https://devfeed.tech/topics/generative-ai.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Image](<https://devfeed.tech/topics/image.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [generative](<https://devfeed.tech/tags/generative.md>), [generative-ai](<https://devfeed.tech/tags/generative-ai.md>), [maps](<https://devfeed.tech/tags/maps.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [planning](<https://devfeed.tech/tags/planning.md>), [satellite](<https://devfeed.tech/tags/satellite.md>)

### AI overview

DualControlNet jointly generates high-resolution land-use maps and satellite imagery from text prompts and multidimensional controls. Trained on more than 40,000 paired tiles from Orlando and Chicago, it is evaluated with land-use, spatial-consistency, and image-quality metrics. Compared with unimodal baselines, it reports 62.8% higher land-use diversity consistency, 16.3% higher building-coverage-ratio consistency, more than twice the cross-modal spatial consistency, and three times the visual quality. Scenario experiments include increasing density and diversity and responding to prompts such as "high-density development." These findings are based on the supplied abstract; it does not provide further methodological detail.

### Source excerpt

arXiv:2611.05859v1 Announce Type: cross Abstract: Various land-use models characterize land-use patterns and facilitate scenario planning. However, existing models are largely predictive, forecasting transitions based on numeric data without synthesizing new land-use maps from simple text prompts. Meanwhile, generative AI in urban systems remains unimodal, weakly controllable, and evaluated mostly by generic image quality rather than domain-specific metrics. To address these gaps, we introduce DualControlNet, a generative land-use model that jointly synthesizes high-resolution land-use maps and satellite imagery, addressing multimodality, controllability, and verification challenges. Trained on over 40,000 paired tiles from Orlando and Chicago, DualControlNet is evaluated using land-use, spatial consistency, and image quality metrics. Relative to unimodal baselines, DualControlNet improves land-use diversity consistency by 62.8% and building coverage ratio consistency by 16.3%, more than doubles cross-modal spatial consistency, and triples visual quality. Scenario-based experiments demonstrate that DualControlNet supports sustainable development through multidimensional controls, increasing density and diversity, and responding to text prompts like "high-density development." Overall, this work advances land-use modeling via a multimodal, controllable, and verifiable generative framework, lowering technical barriers and enabling intuitive exploration of urban futures.

## GeoFaham: Domain-Specialized Multi-Agent Geospatial Analytics for Disaster Response

DevFeed: [GeoFaham: Domain-Specialized Multi-Agent Geospatial Analytics for Disaster Response](<https://devfeed.tech/articles/geofaham-domain-specialized-multi-agent-geospatial-analytics-for-disaster-response-117551.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.04977>)

Author: Muhammad Sohail Danish, Gilles Q. Hacheme, Akram Zaytar, Satej S. Soman, Inbal Becker-Reshef, Juan Lavista Ferres, Salman Khan, Girmaw Abebe Tadesse

Published: 2026-11-04T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Multi-step AI processes](<https://devfeed.tech/topics/multi-step-ai-processes.md>)

Tags: [17](<https://devfeed.tech/tags/17.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [also](<https://devfeed.tech/tags/also.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [applying](<https://devfeed.tech/tags/applying.md>), [artifacts](<https://devfeed.tech/tags/artifacts.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [bench](<https://devfeed.tech/tags/bench.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [best](<https://devfeed.tech/tags/best.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-code](<https://devfeed.tech/tags/claude-code.md>), [code](<https://devfeed.tech/tags/code.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-ma](<https://devfeed.tech/tags/cs-ma.md>), [disaster](<https://devfeed.tech/tags/disaster.md>), [domain](<https://devfeed.tech/tags/domain.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [multi-agent](<https://devfeed.tech/tags/multi-agent.md>)

### AI overview

GeoFaham coordinates specialist agents for vector operations, map feature retrieval, imagery discovery, and raster analysis through an Orchestrator that plans dependencies, tracks progress, validates results, and replans. A Response Store keeps full geospatial outputs and typed metadata outside the LLM context, while agents share compact identifiers and summaries. On GeoFaham-Bench, a 40-task benchmark based on 17 disaster records, it scored 0.940 versus 0.879 for the next-best general-purpose coding-agent baseline, using 7.5 times fewer tokens. The reported evaluation covers multi-step reasoning, cross-source integration, raster-vector computation, artifact generation, and recovery; the evidence here is the abstract, not the full paper.

### Source excerpt

arXiv:2611.04977v1 Announce Type: cross Abstract: Real-world geospatial decision support requires more than single-step retrieval or code generation. Disaster-response questions often involve multi-source workflows that combine spatial databases, map-derived features, satellite imagery, raster products, and reusable geospatial artifacts, where correctness depends on preserving spatial scope, temporal context, and intermediate execution state. While large language models and coding agents offer a promising interface for such workflows, they struggle with geospatial-specific failure modes, including handling large artifacts, selecting the wrong evidence source, losing spatial or temporal metadata, and applying invalid geometric operations. We introduce GeoFaham, a domain-specialized multi-agent system for geospatial disaster analytics. GeoFaham decomposes workflows into specialist agents for vector operations, map feature retrieval, imagery discovery, and raster analysis, coordinated by an Orchestrator with a dual-ledger control loop for dependency-aware planning, progress tracking, validation, and replanning. To support artifact-centric collaboration, GeoFaham uses a Response Store that persists full geospatial outputs and typed metadata outside the LLM context window while agents exchange compact identifiers and summaries. We also introduce GeoFaham-Bench, a 40-task benchmark grounded in 17 real-world disaster records, designed to evaluate multi-step geospatial reasoning, cross-source integration, raster-vector computation, artifact generation, and recovery under non-rigid ground truth. We evaluate GeoFaham on GeoFaham-Bench against strong general-purpose coding-agent baselines, including Codex, Claude Code, and OpenCode. GeoFaham achieves the best overall performance, scoring 0.940 versus the next-best 0.879, while using 7.5 times fewer tokens. Our code and benchmark are publicly available at https://github.com/microsoft/GeoFaham.

## OmniST: Topology-Invariant Spatiotemporal Forecasting with Guaranteed-Coverage Uncertainty at Unseen Sensor Locations

DevFeed: [OmniST: Topology-Invariant Spatiotemporal Forecasting with Guaranteed-Coverage Uncertainty at Unseen Sensor Locations](<https://devfeed.tech/articles/omnist-topology-invariant-spatiotemporal-forecasting-with-guaranteed-coverage-uncertainty-at-unseen-sensor-locations-116922.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.04256>)

Author: Praveg Vashishtha

Published: 2026-11-04T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [generalization in machine learning](<https://devfeed.tech/topics/generalization-in-machine-learning.md>), [Time Series](<https://devfeed.tech/topics/time-series.md>), [model architecture](<https://devfeed.tech/topics/model-architecture.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>)

Tags: [2019](<https://devfeed.tech/tags/2019.md>), [2023](<https://devfeed.tech/tags/2023.md>), [about](<https://devfeed.tech/tags/about.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [added](<https://devfeed.tech/tags/added.md>), [adjacency-matrix](<https://devfeed.tech/tags/adjacency-matrix.md>), [air](<https://devfeed.tech/tags/air.md>), [air-quality](<https://devfeed.tech/tags/air-quality.md>), [also](<https://devfeed.tech/tags/also.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [attention](<https://devfeed.tech/tags/attention.md>), [based](<https://devfeed.tech/tags/based.md>), [coverage](<https://devfeed.tech/tags/coverage.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [invariant](<https://devfeed.tech/tags/invariant.md>), [sensor](<https://devfeed.tech/tags/sensor.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>), [topology](<https://devfeed.tech/tags/topology.md>), [uncertainty](<https://devfeed.tech/tags/uncertainty.md>), [with](<https://devfeed.tech/tags/with.md>)

### AI overview

OmniST forecasts at arbitrary sensor locations by tokenizing coordinates, applying geographic k-nearest-neighbor attention at each forward pass, and processing each node's time series with a bidirectional GRU, avoiding dependence on a fixed node count. Combined with inductive split-conformal calibration, it provides finite-sample coverage guarantees for prediction intervals at unseen locations, including as sensors are added or removed. On Delhi air-quality data, it reports 31.0 µg/m³ 30-day PM2.5 MAE and at most 1.1% degradation after removing 40% of stations. Its unseen-station forecasts outperform nearest-neighbor interpolation and IGNNK in the reported comparisons. Coverage remains at least 90% under 30-50% node dropout across four datasets, while temporal drift causes static conformal coverage to fall to about 0.50 late in the period. A rolling adaptive estimator restores coverage to 0.89 overall and 0.96 late, with wider intervals.

### Source excerpt

arXiv:2611.04256v1 Announce Type: new Abstract: Standard spatiotemporal graph convolutional networks need a fixed-size adjacency matrix, coupling model weights to a specific sensor count N; they cannot produce any output at a sensor absent from training. We present OmniST, a topology-invariant forecaster whose computation has zero structural dependency on the number of nodes, and which, composed with inductive split-conformal calibration, gives distribution-free, coverage-guaranteed prediction intervals at arbitrary (including never-seen) sensor locations, preserving that guarantee as sensors are added or removed. Topology invariance follows from coordinate-based tokenization, k-nearest-neighbor attention from geographic coordinates at each forward pass, and per-node temporal processing via a bidirectional GRU. On real Delhi air-quality data (37 stations, 2019-2023), OmniST attains a 30-day PM2.5 MAE of 31.0 ug/m3 and degrades at most 1.1 percent when 40 percent of stations are removed. Its inductive forecasts at unseen stations beat nearest-neighbor interpolation (about 12 percent) and a learned kriging GNN (IGNNK, about 21 percent), and, unlike either, carry a finite-sample coverage guarantee. The guarantee holds at 90 percent or above under 30-50 percent node dropout across four real datasets over three domains (Delhi and Beijing air quality, NOAA GHCN weather, Caltrans PeMSD7-M traffic), and still holds at 325 real sensors (full PeMS-BAY). We also fix a temporal-drift failure: static conformal under-covers over time (about 0.50 late), but a rolling adaptive-conformal estimator restores target coverage (0.89 overall, 0.96 late) at the cost of wider intervals.

## Arnis turns any place on Earth into a Minecraft world

DevFeed: [Arnis turns any place on Earth into a Minecraft world](<https://devfeed.tech/articles/arnis-turns-any-place-on-earth-into-a-minecraft-world-114845.md>)

Original publisher: [Read original article](<https://github.blog/open-source/maintainers/arnis-turns-any-place-on-earth-into-a-minecraft-world/>)

Author: Alex Perrault

Published: 2026-11-03T16:00:00Z

Content type: article

Language: en

Sources: [GitHub Blog](<https://devfeed.tech/sources/github-engineering.md>)

Topics: [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Code review](<https://devfeed.tech/topics/code-review.md>), [Pull Request](<https://devfeed.tech/topics/pull-request.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [Test-driven development](<https://devfeed.tech/topics/tdd.md>)

Tags: [2022](<https://devfeed.tech/tags/2022.md>), [3d](<https://devfeed.tech/tags/3d.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [app](<https://devfeed.tech/tags/app.md>), [blocks](<https://devfeed.tech/tags/blocks.md>), [bug](<https://devfeed.tech/tags/bug.md>), [bugs](<https://devfeed.tech/tags/bugs.md>), [city](<https://devfeed.tech/tags/city.md>), [classroom](<https://devfeed.tech/tags/classroom.md>), [code](<https://devfeed.tech/tags/code.md>), [code-reviews](<https://devfeed.tech/tags/code-reviews.md>), [coding](<https://devfeed.tech/tags/coding.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [earth](<https://devfeed.tech/tags/earth.md>), [github](<https://devfeed.tech/tags/github.md>), [github-copilot](<https://devfeed.tech/tags/github-copilot.md>), [github-copilot-app](<https://devfeed.tech/tags/github-copilot-app.md>), [maintainers](<https://devfeed.tech/tags/maintainers.md>), [minecraft](<https://devfeed.tech/tags/minecraft.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [open-source-community](<https://devfeed.tech/tags/open-source-community.md>), [prototype](<https://devfeed.tech/tags/prototype.md>), [world](<https://devfeed.tech/tags/world.md>)

### AI overview

Arnis turns selected real-world locations into Minecraft worlds using mapped data from OpenStreetMap and other sources. It places mapped elements according to their coordinates and types, so unmapped features cannot appear; detailed landmarks can instead be built from 3D models. Users choose a place and spawn point, start generation, then can add missing map features to OpenStreetMap. The project began as a Python prototype and was later ported to Rust after open GitHub discussions. Its community contributes features and translations. As sole maintainer, Louis Erbkamm uses GitHub Copilot to triage issues, trace and fix bugs, review large pull requests for performance risks, and check sensitive code paths for regression cases. The excerpt says Arnis has also added the Moon and Mars and describes classroom projects using generated neighborhoods.

### Source excerpt

Louis Erbkamm built the first prototype in his free time. Now GitHub Copilot helps him maintain it alone. The post Arnis turns any place on Earth into a Minecraft world appeared first on The GitHub Blog.

## SpatialVec: A Unified Representation Learning Framework for Geospatial Objects with Boundary-Aware Sampling

DevFeed: [SpatialVec: A Unified Representation Learning Framework for Geospatial Objects with Boundary-Aware Sampling](<https://devfeed.tech/articles/spatialvec-a-unified-representation-learning-framework-for-geospatial-objects-with-boundary-aware-sampling-113639.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.00111>)

Author: Samya Acharja, Kanchan Chowdhury

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Databases](<https://devfeed.tech/sources/arxiv-databases.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Transformers](<https://devfeed.tech/topics/transformers.md>)

Tags: [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cs-cv](<https://devfeed.tech/tags/cs-cv.md>), [cs-db](<https://devfeed.tech/tags/cs-db.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [decoder](<https://devfeed.tech/tags/decoder.md>), [embedding](<https://devfeed.tech/tags/embedding.md>), [encoding](<https://devfeed.tech/tags/encoding.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fine](<https://devfeed.tech/tags/fine.md>), [framework](<https://devfeed.tech/tags/framework.md>), [geospatial](<https://devfeed.tech/tags/geospatial.md>), [learning](<https://devfeed.tech/tags/learning.md>), [local](<https://devfeed.tech/tags/local.md>), [location](<https://devfeed.tech/tags/location.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [modeling](<https://devfeed.tech/tags/modeling.md>), [models](<https://devfeed.tech/tags/models.md>), [points](<https://devfeed.tech/tags/points.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [sampling](<https://devfeed.tech/tags/sampling.md>)

### AI overview

SpatialVec uses boundary-aware sampling to improve geospatial representations of topology-sensitive relationships involving complex shapes. It adaptively samples geometry boundaries and surrounding regions, then uses a transformer encoder-decoder trained through geometry reconstruction to create unified representations for points, polylines, and polygons. The abstract reports evaluation on two real-world datasets, with improved preservation of challenging polygonal topological relationships and significantly better downstream prediction and classification performance than state-of-the-art baselines; it gives no metrics or further methodological detail.

### Source excerpt

arXiv:2611.00111v1 Announce Type: cross Abstract: Spatial representation learning is essential for Geospatial Artificial Intelligence (GeoAI) applications, enabling machine learning models to capture the geometry, location, and spatial relationships of geographic entities such as points, polylines, and polygons. Existing spatial encoding methods often struggle to preserve fine-grained topological information, particularly for interactions involving complex geometries such as polylines and polygons. These approaches inadequately capture boundary structure and local spatial context, causing topology-sensitive relationships near geometry boundaries to be underrepresented in the learned embedding space. Consequently, accurately modeling complex spatial interactions remains a significant challenge for downstream spatial reasoning tasks. To fill the gap, we introduce SpatialVec, a boundary-aware spatial representation learning framework that adaptively samples geometry boundaries and their surrounding regions, capturing fine-grained geometric structures critical for topology-aware spatial reasoning. A transformer-based encoder-decoder trained through geometry reconstruction produces compact, geometry-aware, and unified representations for points, polylines, and polygons. Empirical evaluation on two real-world datasets shows that SpatialVec substantially improves the preservation of challenging topological relationships involving polygonal interactions and outperforms state-of-the-art baselines significantly on downstream prediction and classification tasks.

## Inferring missing locations in human mobility data using memory effects

DevFeed: [Inferring missing locations in human mobility data using memory effects](<https://devfeed.tech/articles/inferring-missing-locations-in-human-mobility-data-using-memory-effects-114525.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.01319>)

Author: Pablo Rosillo-Rodes, Ant\\\`onia Tugores, David S\\'anchez, Jos\\'e J. Ramasco

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Data science and engineering](<https://devfeed.tech/topics/data-science-and-engineering.md>), [geospatial](<https://devfeed.tech/topics/geospatial.md>)

Tags: [bayesian](<https://devfeed.tech/tags/bayesian.md>), [coverage](<https://devfeed.tech/tags/coverage.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [markov-chain](<https://devfeed.tech/tags/markov-chain.md>), [memory](<https://devfeed.tech/tags/memory.md>), [missing-data](<https://devfeed.tech/tags/missing-data.md>), [mobility](<https://devfeed.tech/tags/mobility.md>), [noise](<https://devfeed.tech/tags/noise.md>), [physics-soc-ph](<https://devfeed.tech/tags/physics-soc-ph.md>), [temporal](<https://devfeed.tech/tags/temporal.md>), [twitter](<https://devfeed.tech/tags/twitter.md>)

### AI overview

The proposed method imputes missing locations by weighting observations on both sides of a gap according to their temporal distance, with one decay-rate parameter per individual. Bayesian model comparison selects a power-law decay. On synthetic and real Twitter data, it achieves macro-AUROC above 0.9 overall and above 0.8 for the most mobile users. It matches a gap-aware continuous-time Markov chain overall and outperforms it among the most mobile users. The abstract reports robustness to moderate noise and generalization across countries and European administrative regions.

### Source excerpt

arXiv:2611.01319v1 Announce Type: new Abstract: Empirical datasets used to characterize human mobility are sparse in space and time, owing to technological factors such as coverage and battery, or to the erratic nature of online interactions. Missing data can bias the conclusions drawn from them, so filling the gaps in incomplete timelines is a key challenge in computational social science. Here, we propose a simple imputation method in which the informativeness of an observation decays with its bidirectional temporal distance from the missing entry, at a rate set by a single parameter per individual. On synthetic and real Twitter data, a power-law decay, selected by Bayesian model comparison, yields accurate imputations (macro-AUROC above 0.9 overall, and above 0.8 for the most mobile users) and matches a gap-aware continuous-time Markov chain, outperforming it among the most mobile users. The method is robust to moderate noise and generalizes across countries and European administrative regions.

[Next page](<https://devfeed.tech/topics/geospatial.md?cursor=WyIyMDI2LTExLTAzVDA1OjAwOjAwKzAwOjAwIiwgIjhiOTk3MDcyLWNlOGItNGM2Ni05YTk5LTA1OWQ2ZGNjYzk1ZiJd>)