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

TL;DR

ReSearch addresses the bottleneck of discovering relevant Earth Science datasets by turning broad scientific intent into structured data queries through a multi-stage process. It combines lexical search, semantic embeddings, abbreviation expansion, and large language model reranking in a staged pipeline that separates recall and precision. A literature-grounded benchmark demonstrates that ReSearch outperforms baselines, particularly for task-based, high-level research goals. The framework offers a scalable, interpretable approach to data discovery that can enhance reproducibility and broaden participation in Earth Science research.

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.

Figures (1)

• Figure 1: Evaluation dataset construction pipeline. From academic papers, we extract queries and dataset references, then match datasets to NASA CMR entries to establish ground truth for retrieval evaluation.