Prospective multi-pathogen disease forecasting using autonomous LLM-guided tree search

2026-05-15

AI Summary

This study addresses the limitations of manual modeling in infectious disease forecasting—particularly poor scalability, difficulty adapting to emerging pathogens, and challenges in fine-grained regional prediction—by proposing a large language model–guided tree search framework that autonomously iterates through the generation, evaluation, and optimization of executable forecasting models. This approach achieves, for the first time, autonomous multi-pathogen modeling (influenza, SARS-CoV-2, and RSV), effectively initiating predictions even under data-scarce conditions. Scientific rigor and interpretability are ensured through automated ensemble construction, log-scale distance optimization, and alignment with established epidemiological theory. In real-time prospective forecasts for the 2025–2026 U.S. respiratory virus season, the automatically generated model ensemble matched or surpassed the performance of the CDC’s official human expert ensemble, notably overcoming the cold-start challenge for RSV prediction.

Abstract

Probabilistic forecasting of infectious diseases is crucial for public health but relies on labor-intensive manual model curation by expert modeling teams. This bespoke development bottlenecks scalability to granular geographic resolutions or emerging pathogens. Here, we present an autonomous system using Large Language Model (LLM)-guided tree search to iteratively generate, evaluate, and optimize executable forecasting software. In a fully prospective, real-time evaluation during the 2025-2026 US respiratory season, the system autonomously discovered methodologically diverse models for influenza, COVID-19, and respiratory syncytial virus (RSV). Aggregating these machine-generated models yielded an ensemble that consistently matched or outperformed the gold-standard, human-curated Centers for Disease Control and Prevention (CDC) hub ensembles out-of-sample. The system successfully navigated data-scarce"cold start"scenarios for RSV. Moreover, controlled retrospective ablations revealed that optimizing log-scale distance metrics prevents reward hacking, while an automated judge-in-the-loop ensures structural fidelity to complex scientific theories. By autonomously translating epidemiological theory into accurate, transparent code, this framework overcomes the modeling labor bottleneck, enabling rapid deployment of expert-level disease forecasting at unprecedented scales.

Problem

• infectious disease forecasting
• model curation bottleneck
• scalability
• emerging pathogens
• multi-pathogen

Innovation

• LLM-guided tree search
• autonomous forecasting
• multi-pathogen prediction
• cold start modeling
• scientific code generation