ReSearch: A Multi-Stage Machine Learning Framework for Earth Science Data Discovery

2026-01-20

AI Summary

This work addresses the growing challenge in Earth science where explosive data growth hinders effective alignment between research objectives and heterogeneous datasets, as existing retrieval systems struggle to interpret high-level scientific intent. To bridge this gap, we propose a multi-stage, reasoning-enhanced data discovery framework that maps scientific questions to relevant datasets through a three-phase pipeline: intent parsing, high-recall retrieval, and context-aware reranking. Our approach innovatively integrates intent understanding, multi-stage retrieval, and large language model–based reranking, explicitly decoupling recall and precision objectives. We also construct the first evaluation benchmark derived from real scientific literature. Experimental results demonstrate that our method significantly outperforms baseline approaches on task-oriented queries, achieving consistent improvements in both recall and ranking performance—particularly for abstractly formulated scientific goals.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Search and Optimization: Distributed Search

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Web Mining and Content Analysis: Interdisciplinary science discovery with web data mining

Abstract

The rapid expansion of Earth Science data from satellite observations, reanalysis products, and numerical simulations has created a critical bottleneck in scientific discovery, namely identifying relevant datasets for a given research objective. Existing discovery systems are primarily retrieval-centric and struggle to bridge the gap between high-level scientific intent and heterogeneous metadata at scale. We introduce \textbf{ReSearch}, a multi-stage, reasoning-enhanced search framework that formulates Earth Science data discovery as an iterative process of intent interpretation, high-recall retrieval, and context-aware ranking. ReSearch integrates lexical search, semantic embeddings, abbreviation expansion, and large language model reranking within a unified architecture that explicitly separates recall and precision objectives. To enable realistic evaluation, we construct a literature-grounded benchmark by aligning natural language intent with datasets cited in peer-reviewed Earth Science studies. Experiments demonstrate that ReSearch consistently improves recall and ranking performance over baseline methods, particularly for task-based queries expressing abstract scientific goals. These results underscore the importance of intent-aware, multi-stage search as a foundational capability for reproducible and scalable Earth Science research.

Problem

• Earth Science data discovery
• scientific intent
• heterogeneous metadata
• data retrieval bottleneck
• dataset identification

Innovation

• multi-stage search
• intent-aware retrieval
• semantic embedding
• large language model reranking
• Earth Science data discovery