ChinaTravel: An Open-Ended Travel Planning Benchmark with Compositional Constraint Validation for Language Agents
ChinaTravel: 一个带有组合约束验证的开放旅行规划基准,用于语言代理
机构 * State Key Laboratory of Novel Software Technology, Nanjing University, China(南京大学新型软件技术国家重点实验室) ; School of Intelligence Science and Technology, Nanjing University, China(南京大学智能科学与技术学院) ; School of Artificial Intelligence, Nanjing University, China(南京大学人工智能学院) ; Noah’s Ark Lab, Huawei, China(华为诺亚实验室)
AI总结 ChinaTravel通过开放数据集和组合通用领域语言,提升语言代理在复杂真实场景中的约束满足能力,实现37%的约束满足率,比纯神经模型提升10倍。
Comments ICLR 2026. Webpage: https://www.lamda.nju.edu.cn/shaojj/chinatravel