GeoDecider: A Coarse-to-Fine Agentic Workflow for Explainable Lithology Classification

2026-05-05

AI Summary

This work proposes GeoDecider, a training-free large language model agent framework that reframes lithology classification as a coarse-to-fine, multi-stage expert reasoning process. Unlike conventional approaches that treat lithology classification as a single-step prediction task and lack geological interpretability, GeoDecider integrates geological prior knowledge and external tools through structured reasoning stages: initial coarse categorization, tool-augmented inference leveraging pretrained classifiers, contextual analysis, and neighborhood retrieval, followed by geologically consistent post-processing. By embedding expert-like reasoning and domain knowledge without requiring model fine-tuning, GeoDecider achieves state-of-the-art performance across four benchmark datasets, simultaneously delivering high accuracy, computational efficiency, and geological interpretability.

Abstract

Lithology classification aims to infer subsurface rock types from well-logging signals, supporting downstream applications like reservoir characterization. Despite substantial progress, most existing methods still treat lithology classification as a single-pass classification task. In contrast, practical experts incorporate geological principles, external knowledge, and tool-use capabilities to perform accurate classification. In this work, we propose GeoDecider, a coarse-to-fine agentic workflow that enables accurate and explainable lithology classification through training-free use of large language models (LLMs). GeoDecider reformulates lithology classification as an expert-like structured process and organizes it into a multi-stage workflow involving coarse-to-fine reasoning. Specifically, GeoDecider includes the following stages: (1) base classifier-guided coarse classification, which uses a pre-trained classifier to provide a rough reference for downstream tasks, thus reducing the overall cost of downstream reasoning, (2) tool-augmented reasoning, which utilizes several tools such as contextual analysis and neighbor retrieval to achieve finer and more precise classifications, (3) geological refinement, which post-processes the final results to enforce geological consistency. Experiments on four benchmarks show that GeoDecider outperforms representative baselines. Further analysis demonstrates that the proposed framework produces geologically interpretable predictions while achieving a better trade-off between classification performance and inference efficiency.

Problem

• lithology classification
• explainable AI
• geological consistency
• well-logging signals
• expert-like reasoning

Innovation

• coarse-to-fine reasoning
• agentic workflow
• explainable lithology classification
• tool-augmented reasoning
• large language models