Back to home返回首页
QUESTIONS I KEEP COMING BACK TO
Research in progress研究进行时.
Coordination, information, and behavior in intelligent mobility systems.智能出行系统中的协调、信息与行为。
CURRENT QUESTIONS当前问题
Current questions正在探索的问题
Information & autonomous decisions信息与自主决策
Personal history & behavioral intelligence个人历史与行为智能
A CONNECTED RESEARCH AGENDA相互连接的研究脉络
Learning. Decisions. Coordination.学习、决策与协调
Learning & Reasoning学习与推理
How can agents learn useful representations and reason about complex environments?智能主体如何学习有效表征，并对复杂环境进行推理？
Representations · Knowledge · ReasoningMulti-Agent Decision-Making多智能体决策
How should agents make decisions when their actions and objectives interact?当行为相互影响、目标各不相同时，智能主体应如何决策？
Reinforcement learning · Game theoryCoordination & Collective Intelligence协调与集体智能
How can interaction among agents lead to effective collective behavior?多个智能主体如何通过互动，形成有效的集体行为？
Cooperation · Interaction · Collective behaviorFoundations & earlier work研究基础与已有工作
Papers & manuscripts论文与研究稿件Understanding travel behavior from trajectories从轨迹理解出行行为
My master's work explored travel intention and behavioral representation using mobility traces and language models.硕士阶段的研究围绕出行轨迹与语言模型，探索出行意图和行为表征。
Trajectory learning · Behavioral representation · LLM reasoningEvent-CausNet: Unlocking Causal Knowledge from Text with LLMs for Reliable Spatio-Temporal Forecasting
Turning event reports into causal knowledge for more reliable traffic forecasting.将事件文本转化为因果知识，支持更可靠的交通预测。
Paper & resources论文与资源AskNearby: An LLM-based Application for Neighborhood Information Retrieval and Personalized Cognitive-map Recommendations
Connecting neighborhood knowledge with the places and preferences that matter to each person.结合社区知识、地点与个人偏好，让身边的信息更易获取。
Paper & resources论文与资源ST-ProC: A Graph-Prototypical Framework for Robust Semi-Supervised Travel Mode Identification
Learning robust travel-mode representations from trajectories with limited labels.在标签稀缺的条件下，从轨迹中学习稳健的出行方式表征。
Paper & resources论文与资源MF-AttnBiLSTM: Traffic Flow Prediction via Hybrid Signal Decomposition and Dual-Stream Temporal Attention Learning
Paper & resources论文与资源Routing & multi-agent systems路径决策与多智能体系统
Earlier projects studied traffic assignment and coordination in unmanned warehouses, connecting individual decisions with shared system constraints.此前的项目涉及交通分配与无人仓协同，关注个体决策与系统共同约束之间的联系。
Network optimization · Multi-agent scheduling