@inproceedings{tang-etal-2026-geogr2,
title = "{G}eo{GR}{\texttwosuperior}: Zero-Shot Geospatial Inference via Geostatistically-Guided Iterative Refinement with {LLM}s",
author = "Tang, Jinfan and
Wu, Kunming and
Gong, Xieruifeng and
He, Yuya and
Jie, Hu and
Junhui, Wu and
Wu, Yuankai",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.310/",
pages = "6560--6580",
ISBN = "979-8-89176-462-0",
abstract = "Standard large language model prompting treats geospatial inference as independent, instance-wise prediction, ignoring the fundamental spatial dependencies that govern geographic reality. Consequently, even advanced models struggle with spatial consistency and exhibit severe biases toward populous regions. To bridge this gap, we propose **GeoGR{\texttwosuperior}** (**Geo**spatial **G**raph **R**efine **R**easoning), a framework that formalizes zero-shot geospatial prediction as an iterative message-passing process on a dynamically constructed graph. Unlike static retrieval methods, GeoGR{\texttwosuperior} instantiates three dynamic operators via collaborating operators: (1) a **Topology Operator** that constructs graph topology to enforce the Spatial Markov property; (2) a **Feature Operator** that enriches nodes with task-relevant semantic covariates; and (3) an **Update Operator** that performs natural language message passing to iteratively minimize spatial discrepancy. Theoretically, we frame this refinement as a contraction mapping that approximates the fixed point of a global consistency equation. Empirically, we validate GeoGR{\texttwosuperior} on diverse physical and socioeconomic tasks. Results demonstrate that by explicitly embedding geostatistical inductive biases, GeoGR{\texttwosuperior} significantly outperforms standard prompting baselines, while effectively mitigating systematic geographic bias. Our framework leverages large language models' intrinsic capacity for understanding spatial correlations through explicit topological scaffolding, without resorting to general graph reasoning paradigms."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="tang-etal-2026-geogr2">
<titleInfo>
<title>GeoGR²: Zero-Shot Geospatial Inference via Geostatistically-Guided Iterative Refinement with LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jinfan</namePart>
<namePart type="family">Tang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kunming</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xieruifeng</namePart>
<namePart type="family">Gong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuya</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hu</namePart>
<namePart type="family">Jie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wu</namePart>
<namePart type="family">Junhui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuankai</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Standard large language model prompting treats geospatial inference as independent, instance-wise prediction, ignoring the fundamental spatial dependencies that govern geographic reality. Consequently, even advanced models struggle with spatial consistency and exhibit severe biases toward populous regions. To bridge this gap, we propose **GeoGR²** (**Geo**spatial **G**raph **R**efine **R**easoning), a framework that formalizes zero-shot geospatial prediction as an iterative message-passing process on a dynamically constructed graph. Unlike static retrieval methods, GeoGR² instantiates three dynamic operators via collaborating operators: (1) a **Topology Operator** that constructs graph topology to enforce the Spatial Markov property; (2) a **Feature Operator** that enriches nodes with task-relevant semantic covariates; and (3) an **Update Operator** that performs natural language message passing to iteratively minimize spatial discrepancy. Theoretically, we frame this refinement as a contraction mapping that approximates the fixed point of a global consistency equation. Empirically, we validate GeoGR² on diverse physical and socioeconomic tasks. Results demonstrate that by explicitly embedding geostatistical inductive biases, GeoGR² significantly outperforms standard prompting baselines, while effectively mitigating systematic geographic bias. Our framework leverages large language models’ intrinsic capacity for understanding spatial correlations through explicit topological scaffolding, without resorting to general graph reasoning paradigms.</abstract>
<identifier type="citekey">tang-etal-2026-geogr2</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.310/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>6560</start>
<end>6580</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T GeoGR²: Zero-Shot Geospatial Inference via Geostatistically-Guided Iterative Refinement with LLMs
%A Tang, Jinfan
%A Wu, Kunming
%A Gong, Xieruifeng
%A He, Yuya
%A Jie, Hu
%A Junhui, Wu
%A Wu, Yuankai
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F tang-etal-2026-geogr2
%X Standard large language model prompting treats geospatial inference as independent, instance-wise prediction, ignoring the fundamental spatial dependencies that govern geographic reality. Consequently, even advanced models struggle with spatial consistency and exhibit severe biases toward populous regions. To bridge this gap, we propose **GeoGR²** (**Geo**spatial **G**raph **R**efine **R**easoning), a framework that formalizes zero-shot geospatial prediction as an iterative message-passing process on a dynamically constructed graph. Unlike static retrieval methods, GeoGR² instantiates three dynamic operators via collaborating operators: (1) a **Topology Operator** that constructs graph topology to enforce the Spatial Markov property; (2) a **Feature Operator** that enriches nodes with task-relevant semantic covariates; and (3) an **Update Operator** that performs natural language message passing to iteratively minimize spatial discrepancy. Theoretically, we frame this refinement as a contraction mapping that approximates the fixed point of a global consistency equation. Empirically, we validate GeoGR² on diverse physical and socioeconomic tasks. Results demonstrate that by explicitly embedding geostatistical inductive biases, GeoGR² significantly outperforms standard prompting baselines, while effectively mitigating systematic geographic bias. Our framework leverages large language models’ intrinsic capacity for understanding spatial correlations through explicit topological scaffolding, without resorting to general graph reasoning paradigms.
%U https://aclanthology.org/2026.findings-emnlp.310/
%P 6560-6580
Markdown (Informal)
[GeoGR²: Zero-Shot Geospatial Inference via Geostatistically-Guided Iterative Refinement with LLMs](https://aclanthology.org/2026.findings-emnlp.310/) (Tang et al., Findings 2026)
ACL