Peilong Ma

Optimizing AI-driven Geographic Simulation Task Scheduling through Intelligent Runtime Estimation for Distributed Heterogeneous Clusters

Tue, 10 Mar 2026 00:00:00 +0000

This under-review work studies how runtime estimation and resource-aware scheduling can improve throughput and balance for AI-driven geo-simulation tasks on distributed heterogeneous clusters.

PyGeoModel: A Python Package for Integrating Intelligent Geographic Model Services for Urban Analysis in Jupyter

Tue, 10 Mar 2026 00:00:00 +0000

This under-review paper introduces PyGeoModel as a Jupyter-oriented interface for discovering, configuring, and executing intelligent geographic model services in urban analysis workflows.

Chat2Map: A ReAct-based Agent Framework for Automated Web Map Generation from Natural Language Instructions

Mon, 16 Feb 2026 00:00:00 +0000

This accepted conference paper introduces Chat2Map, a ReAct-based agent framework for translating natural language instructions into executable web map generation workflows with stronger reliability for WebGIS tasks.

OpenGeoLab: Synergizing service-oriented resources and reproducible workflows for geographic modeling

Wed, 14 Jan 2026 00:00:00 +0000

This preprint presents OpenGeoLab as a reference implementation for a unified geographic modeling infrastructure that connects distributed resources, interactive configuration, and reproducible execution environments.

OpenGeoLab: Synergizing service-oriented resources and reproducible workflows for geographic modeling

Sat, 01 Nov 2025 09:00:00 +0800

Presented in Adelaide in November 2025 as part of ongoing work on reproducible geographic modeling workflows.

Facilitating sensitivity analysis of hydrological models through knowledge-driven configuration and distributed online model services

Wed, 01 Oct 2025 00:00:00 +0000

This work studies how configuration knowledge and service-oriented model execution can be combined to support more efficient sensitivity analysis for hydrological models.

PyGeoModel: A Python Package for Integrating Intelligent Geographic Model Services into Jupyter's Interactive Computing Environment

Sun, 01 Jun 2025 09:00:00 +0800

Presented at the 2025 conference in Jiaozuo and received the Best Paper award.

GeoDataProcessor-MCP

Fri, 18 Apr 2025 00:00:00 +0000

GeoDataProcessor-MCP is an MCP server for geospatial data processing tools.

It exposes WhiteBox and SAGA GIS style processing capabilities through a standardized interface so large language models and desktop agents can discover, inspect, and invoke geographic analysis tools.

Knowledge- and model service-based sensitivity analysis for hydrological modeling

Tue, 01 Apr 2025 09:00:00 +0800

Presented in Hangzhou in April 2025 as an early exploration of intelligent geographic modeling.

PyGeoModel

Wed, 01 Jan 2025 00:00:00 +0000

PyGeoModel is an intelligent Python package for urban and geographic modeling within Jupyter environments.

It combines model recommendation, interactive configuration, question answering, and service-based execution to make geographic modeling more accessible in computational notebooks.

RS_Segmentation

Wed, 01 Jan 2025 00:00:00 +0000

RS_Segmentation is an intelligent segmentation system for remote sensing and GIS imagery.

The system supports text-prompt segmentation, point-based refinement, GeoTIFF overlay browsing, Shapefile export, and optional voice interaction for geospatial image interpretation workflows.

Projects

Sun, 19 May 2024 00:00:00 +0000

Construction of an open knowledge framework for geoscientific models

Wed, 10 Jan 2024 00:00:00 +0000

The paper presents a three-level framework covering model resources, their related materials, and practical application knowledge, with OpenGMS as an implementation case.

CV

Tue, 24 Oct 2023 00:00:00 +0000

Retrieval of Fine-Grained PM2.5 Spatiotemporal Resolution Based on Multiple Machine Learning Models

Wed, 26 Jan 2022 00:00:00 +0000

This article evaluates multiple machine learning algorithms for PM2.5 estimation and shows how multi-source urban data can improve fine-grained spatiotemporal retrieval.