From Indicators to Insights: Diversity-Optimized for Medical Series-Text Decoding via LLMs

Jing Wang (Huawei Technologies Ltd.) · Youfang Lin (Beijing jiaotong univercity) · Xiaojun Ning (Beijing Jiaotong University) · Xiyuan Jin (Beijing Jiaotong University) · Ziwei Lin (Beijing Jiaotong University) · QIANRU JIA (BEIJING JIAOTONG UNIVERSITY) · Yuqing Huang (University of Science and Technology of China) · Zhonghua Shi (Capital Medical University )
biomedical literatureclinical guidelinescomputational health analyticscontextual knowledgedecision-making indicatorsevolutionary learningexpert-defined indicatorsiterative optimizationknowledge-aware frameworkmedical benchmarksmedical time-series analysisperformance evaluationprompt engineeringsignal interpretation

Medical time-series analysis differs fundamentally from general ones by requiring specialized domain knowledge to interpret complex signals and clinical context. Large language models (LLMs) hold great promise for augmenting medical time-series analysis by complementing raw series with rich contextual knowledge drawn from biomedical literature and clinical guidelines. However, realizing this potential depends on precise and meaningful prompts that guide the LLM to key information. Yet, determining what constitutes effective prompt content remains non-trivial—especially in medical settings where signal interpretation often hinges on subtle, expert-defined decision-making indicators. To this end, we propose InDiGO, a knowledge-aware evolutionary learning framework that integrates clinical signals and decision-making indicators through iterative optimization. Across four medical benchmarks, InDiGO consistently outperforms prior methods. The code is available at: https://github.com/jinxyBJTU/InDiGO.