GeoColab: an LLM-based multi-agent collaborative framework for geospatial code generation

Automated geospatial code generation using large language models (LLMs) faces challenges in requirement parsing, syntax adaptation, path retrieval, code validation, and spatial recognition, often leading to ‘code hallucinations.’ To address these, we introduce GeoColab, the first multi-agent framewo...

Descrición completa

Detalles Bibliográficos

Principais autores:	Huayi Wu, Haoyue Jiao, Shuyang Hou, Jianyuan Liang, Zhangxiao Shen, Anqi Zhao, Yaxian Qing, Fengying Jin, Xuefeng Guan, Zhipeng Gui
Formato:	Artigo
Idioma:	Inglês
Publicado:	Taylor & Francis Group 2025-12-01
Series:	International Journal of Digital Earth
Assuntos:	Geospatial code generation large language models multi-agent collaboration deepseek evaluation benchmark
Acceso en liña:	https://www.tandfonline.com/doi/10.1080/17538947.2025.2569405
Tags:	Engadir etiqueta Sen Etiquetas, Sexa o primeiro en etiquetar este rexistro!