KDASO：知识-数据双驱动的责任判定技能自动优化
原标题：KDASO: Knowledge-Data Dual-Driven Automated Skill Optimization for Liability Adjudication Tasks
AI 摘要
该论文提出 KDASO 方法，通过知识-数据双驱动自动优化责任判定任务中的技能库。方法包含基于近期判定数据的邻域条件归纳、基于 LLM 轨迹分析的执行对齐精炼，以及上下文预算感知的路由选择。在网约车纠纷、法律责任判定和逻辑推理任务上，KDASO 显著优于通用与领域基线，跨概念相似场景迁移提升 8.96%。
正文节选
KDASO: Knowledge-Data Dual-Driven Automated Skill Optimization for Liability Adjudication Tasks Abstract Liability adjudication tasks (e.g., legal charge determination) require analyzing event descriptions against domain-specific skills to determine whether a party should bear liability. While Large Language Models (LLMs) have shown promise in this area by leveraging executable skills for interpretable reasoning, existing approaches face three critical limitations: skill libraries become stale a