Automated Analysis Framework for Multilingual Climate-Health Literature Based on Multi-Agent Large Language Model

Yuze Sun; Shihui Zhang; Jiancheng Pan; Yunjia Ye; Wentao Luo; Jiahao Li; Quan Zhang; Wenjia Cai

cs.AIAug 28, 2026

Automated Analysis Framework for Multilingual Climate-Health Literature Based on Multi-Agent Large Language Model

Authors: Yuze Sun, Shihui Zhang, Jiancheng Pan, Yunjia Ye, Wentao Luo, Jiahao Li, Quan Zhang, Wenjia Cai,

Organizations: Department of Earth System Science, Ministry of Education Key Laboratory for Earth System Modelling, Institute for Global Change Studies, Tsinghua University, Beijing, China · Huawei Technologies Co., Ltd · Renmin University of China, Beijing, China · School of Information and AI, Beijing Forestry University, Beijing, China ·

Abstract

The rapid proliferation of interdisciplinary and multilingual scientific literature has left traditional manual analysis and single-algorithm methods plagued by low efficiency, poor scalability, and insufficient domain adaptability. Targeting the literature analysis needs of the typical interdisciplinary climate-health field, this study proposes a multi-agent large language model automated analysis framework for multilingual scientific literature, which realizes full-process automation covering literature screening, structured information extraction, and standardized integration. With a central coordination module as the core, the framework deploys three dedicated agents for document evaluation, information extraction, and analytical review to mimic the literature analysis thinking of domain experts, and adopts a four-layer hallucination control strategy together with a manual verification procedure to ensure the accuracy and reliability of analytical outcomes. Validated on a bilingual Chinese-English corpus of 32,642 climate-health papers covering China from 1993 to 2023, the framework achieves an F1 score of 0.92 in core information extraction, and completes the extraction and standardization of 2,012 city-literature association pairs, offering effective technical support for large-scale evidence mining in the climate-health research domain.

Figures & tables

Figure 1: Proposed multi-agent LLM framework for automated climate-health literature analysis. It includes three core agents (Document Appraiser, Information Extractor, and Senior Analyst) supported by external databases and foundational models, enabling end-to-end processing from multi-source literature input to structured data and analysis reports.

Figure 2: Overview of spatial and thematic patterns in China’s climate-health research. (a) City-level actual risk growth rate across China. ( Cai et al. 2025 ) (b) Research hotspots automatically identified by the Senior Analyst framework. (c) Circular heatmap of climate-health research topics, illustrating the relative attention to four categories: health outcomes, climate hazards, vulnerable groups, and socioeconomic determinants, as automatically summarized by the Senior Analyst.

Extraction Task	Precision (P) $\uparrow$	Recall (R) $\uparrow$	F1-Score (F1) $\uparrow$
Geographic Information (City)	0.95	0.93	0.94
Climate Exposure Factor	0.92	0.91	0.91
Health Outcome	0.93	0.92	0.92
Overall Average	0.93	0.92	0.92

Table 1: Performance Comparison of the Proposed Framework on Core Information Extraction Tasks

Method	Precision (P) $\uparrow$	Recall (R) $\uparrow$	F1-Score (F1) $\uparrow$
Single LLM (full merged prompt)	0.84 $\pm$ 0.02	0.81 $\pm$ 0.03	0.82 $\pm$ 0.02
BERT - based domain NER ( Devlin et al. 2019 )	0.86 $\pm$ 0.03	0.83 $\pm$ 0.02	0.84 $\pm$ 0.01
AutoGen (sequential multi - agent) ( Wu et al. 2023 )	0.88 $\pm$ 0.02	0.86 $\pm$ 0.02	0.87 $\pm$ 0.02
LiRA (Literature Review Agents) ( Go et al. 2026 )	0.89 $\pm$ 0.01	0.87 $\pm$ 0.02	0.88 $\pm$ 0.03
Paper Circle ( Kumar et al. 2026 )	0.91 $\pm$ 0.02	0.90 $\pm$ 0.01	0.90 $\pm$ 0.02
Ours (full modular agent pipeline)	0.93 $\pm$ 0.01	0.92 $\pm$ 0.01	0.92 $\pm$ 0.01

Table 2: Baseline comparison on three core extraction tasks and overall performance

Variant	Precision (P) $\uparrow$	Recall (R) $\uparrow$	F1-Score (F1) $\uparrow$
Single - agent (all tasks in one agent)	0.84 $\pm$ 0.02	0.81 $\pm$ 0.03	0.83 $\pm$ 0.02
Two - agent (appraiser + merged extractor - analyst)	0.89 $\pm$ 0.02	0.88 $\pm$ 0.02	0.88 $\pm$ 0.02
Three - agent (our full pipeline)	0.93 $\pm$ 0.01	0.92 $\pm$ 0.01	0.92 $\pm$ 0.01

Table 3: Ablation study of agent architectures with different task - decomposition granularities

Variant	Precision	Recall	F1 - score
Ours (full system)	0.93 $\pm$ 0.01	0.92 $\pm$ 0.01	0.92 $\pm$ 0.01
w/o central coordination module	0.91 $\pm$ 0.02	0.89 $\pm$ 0.02	0.90 $\pm$ 0.02
w/o cross - agent verification	0.89 $\pm$ 0.02	0.88 $\pm$ 0.01	0.88 $\pm$ 0.02
w/o confidence - score filtering	0.87 $\pm$ 0.01	0.90 $\pm$ 0.02	0.88 $\pm$ 0.01
w/o external knowledge - base filtering	0.88 $\pm$ 0.02	0.87 $\pm$ 0.02	0.87 $\pm$ 0.02

Table S1: Ablation on hallucination - control components and central coordination module. Mean

\pm

std over three runs.

LLM Information Extraction

Large language model information extraction converts unstructured documents into structured entities, relations, events, and records.

See all 238 papers

ClimAgent: LLM as Agents for Autonomous Open-ended Climate Science Analysis
Apr 18, 2026Hao Wang, Jindong Han, Wei Fan +1AI Agents for Scientific Discovery Climate Science
1The Hong Kong University of Science and Technology (Guangzhou) · 3Shandong University · University of Auckland +1
CLIMATEAGENT: Multi-Agent Orchestration for Complex Climate Data Science Workflows
Nov 25, 2025Chenyue Li, Hyeonjae Kim, Wen Deng +4AI for Science Multi-Agent Orchestration
1The Hong Kong University of Science and Technology
Large language model-assisted discovery of cohorts from scientific literature
Aug 16, 2026Moritz Sturm, Lisa M. Berg, Inken Berg +6Biomedical IR
Department of Child and Adolescent Psychiatry, Psychosomatic Medicine and Psychotherapy, University Hospital Frankfurt, Goethe University Frankfurt, Frankfurt am Main, Germany · Department of Computer Science, Goethe University Frankfurt, Frankfurt am Main, Germany · The Hessian Center for Artificial Intelligence (hessian.AI), Darmstadt, Germany

Apr 18, 2026J/K move · Enter open · S save

ClimAgent: LLM as Agents for Autonomous Open-ended Climate Science Analysis

Hao Wang, Jindong Han, Wei Fan, Hao Liu

1The Hong Kong University of Science and Technology (Guangzhou) · 3Shandong University · University of Auckland+1

AI Agents for Scientific Discovery Climate Science LLM Agents Large Language Model-Guided Scientific Discovery

Open paper AskPDF arXiv

Climate research is pivotal for mitigating global environmental crises, yet the accelerating volume of multi-scale datasets and the complexity of analytical tools have created significant bottlenecks, constraining scientific discovery to fragmented and labor-intensive workflows. While the emergence Large Language Models (LLMs) offers a transformative paradigm to scale scientific expertise, existing explorations remain largely confined to simple Question-Answering (Q&A) tasks. These approaches often oversimplify real-world challenges, neglecting the intricate physical constraints and the data-driven nature required in professional climate science.To bridge this gap, we introduce ClimAgent, a general-purpose autonomous framework designed to execute a wide spectrum of research tasks across diverse climate sub-fields. By integrating a unified tool-use environment with rigorous reasoning protocols, ClimAgent transcends simple retrieval to perform end-to-end modeling and analysis. To foster systematic evaluation, we propose ClimaBench, the first comprehensive benchmark for real-world climate discovery. It encompasses challenging problems spanning 5 distinct task categories derived from professional scenarios between 2000 and 2025. Experiments on ClimaBench demonstrate that ClimAgent significantly outperforms state-of-the-art baselines, achieving a 40.21% improvement over original LLM solutions in solution rigorousness and practicality. Our code are available at https://github.com/usail-hkust/ClimAgent.

Abstract

LLM Information Extraction

Explore similar work

ClimAgent: LLM as Agents for Autonomous Open-ended Climate Science Analysis

CLIMATEAGENT: Multi-Agent Orchestration for Complex Climate Data Science Workflows

Large language model-assisted discovery of cohorts from scientific literature