Fine-Tuning Small Language Models (SLMs) for Autonomous Web-based Geographical Information Systems (AWebGIS)

2025-08-06

AI Summary

Existing autonomous WebGIS (AWebGIS) systems rely on cloud-based large language models (LLMs), raising concerns regarding privacy leakage, network dependency, and poor scalability. Method: This paper proposes the first fully localized, browser-side AWebGIS solution. It employs a fine-tuned T5-small model deployed entirely on the client to perform high-accuracy semantic parsing of natural language queries into geospatial operation commands, enabling offline inference without any cloud LLM invocation. Contribution/Results: Experiments demonstrate substantial improvements over traditional classification models and cloud-based LLM baselines, achieving exact-match accuracy of 0.93, Levenshtein similarity of 0.99, and ROUGE-1 and ROUGE-L scores of 0.98 each. This work is the first to empirically validate the feasibility of lightweight models for high-fidelity geospatial semantic understanding directly in the browser, establishing a new paradigm for intelligent, privacy-preserving, and low-bandwidth geographic interaction.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Autonomous web-based geographical information systems (AWebGIS) aim to perform geospatial operations from natural language input, providing intuitive, intelligent, and hands-free interaction. However, most current solutions rely on cloud-based large language models (LLMs), which require continuous internet access and raise users' privacy and scalability issues due to centralized server processing. This study compares three approaches to enabling AWebGIS: (1) a fully-automated online method using cloud-based LLMs (e.g., Cohere); (2) a semi-automated offline method using classical machine learning classifiers such as support vector machine and random forest; and (3) a fully autonomous offline (client-side) method based on a fine-tuned small language model (SLM), specifically T5-small model, executed in the client's web browser. The third approach, which leverages SLMs, achieved the highest accuracy among all methods, with an exact matching accuracy of 0.93, Levenshtein similarity of 0.99, and recall-oriented understudy for gisting evaluation ROUGE-1 and ROUGE-L scores of 0.98. Crucially, this client-side computation strategy reduces the load on backend servers by offloading processing to the user's device, eliminating the need for server-based inference. These results highlight the feasibility of browser-executable models for AWebGIS solutions.

Problem

• Enabling autonomous GIS with natural language input
• Reducing reliance on cloud-based LLMs for privacy
• Improving accuracy with client-side SLMs in browsers

Innovation

• Fine-tuned T5-small SLM for client-side processing
• Offline AWebGIS with browser-executable SLM
• Client-side computation reduces server load