Xiaohongshu
ORCID
Hi! I am Kaizhen Tan (Chinese name: 谭楷蓁), an incoming Ph.D. student in Public Administration at New York University, currently pursuing my master’s degree in Artificial Intelligence at Carnegie Mellon University. I received my bachelor’s degree in Information Systems from Tongji University.
My research sits at the intersection of Urban Science and Human-centered AI. Driven by the vision of harmonizing artificial intelligence with urban ecosystems, I aim to build spatially intelligent and socially aware systems that make cities more adaptive, inclusive, and human-centric.
To realize this vision, my work integrates:
- Paradigms: Robot-Friendly City, Self-evolving Urban Digital Twins, Lightweight Urbanization
- Methodologies: Multimodal Learning, Geospatial & Spatiotemporal Data Analysis, Computational Social Science
- Technologies: LLMs, VLMs, AI Agents, Embodied AI, Spatial Intelligence, World Models, Emerging Devices
Specifically, my research agenda explores four key topics:
🤖 1. Robotic Urbanism & Governance
How should embodied intelligence operate in cities, and how can humans govern it at scale?
- Robot-Friendly Urban Space: Redesign streets and buildings with standards for robot siting, infrastructure, and responsibility boundaries.
- Embodied Operation: Vision-based navigation and task execution with accessibility-aware routing under real urban constraints.
- Low-Altitude Governance: Derive air corridors from demand signals and validate them in 3D city models for privacy, noise, and safety.
- Emerging Devices & Deployment: Study city-scale deployment of robots, drones, wearables, and BCI-like devices.
- Public Acceptance & Ethics: Model public perception of risks and capabilities to guide interaction design and rollout strategy.
🏙️ 2. Urban Digital Twins & Agents
How to build a continuously updated digital twin that hosts agents and supports equitable city governance?
- Urban Foundation Model: Fuse remote sensing, street imagery, trajectories, POI, IoT, and text into unified urban representations.
- Measurement & Sensing: Scalable metrics and updating workflows using robots, drones, and wearables for continuous urban sensing.
- Mapping for 3D City: Geo-localization and semantic mapping across point clouds, meshes, and 3D Gaussians for querying and simulation.
- Urban Agents: Task agents for planning and public services — map-LLM, spatial RAG, policy QA, and travel assistance.
- Policy Sandbox: What-if simulation, risk assessment, and policy execution checks within the twin.
🎨 3. Multimodal Social Sensing
How can multimodal human-centered data become actionable insights for urban planning?
- AI-Enhanced Geospatial Analysis: Link urban form, environment, and mobility with human behavior and public service outcomes.
- Accessibility & Pedestrian Experience: Analyze walking experiences and barriers, integrating mobility needs of disabled people into governance.
- Urban Perception & Visual Aesthetics: Quantify streetscape aesthetics and neighborhood imagery to inform design and regeneration priorities.
- Socio-Cultural Signals: Embed place-based narratives into LLM-enabled tools for communication and inclusive governance.
🚀 4. Spatial Intelligence & World Models
How can world models support reliable spatial reasoning and decision-making for physical agents?
- World Models & Architectures: Generative, predictive, and representation-learning architectures for forecasting and planning in the physical world.
- Embodied Representations: Unify geometry, semantics, physics, and action into shared representations for embodied modalities.
- Lifelong Memory & Self-Evolution: Learning mechanisms with stability, forgetting control, and safety constraints for long-horizon autonomy.
- Interpretable Spatial Reasoning: Improve interpretability and robustness via 3D-aware encoders and alternatives to standard transformers.
🔥 News
- 2026.03: 🎓 I am pleased to share that I will begin my PhD at New York University in Fall 2026 under the supervision of Prof. Chenghe Guan and Prof. Zhan Guo.
- 2026.01: 🎉 The abstract co-authored with Prof. Fan Zhang has been accepted for the XXV ISPRS Congress 2026. See you in Toronto!
- 2025.12: 🎉 Our paper, led by my senior labmate Dr. Weihua Huan and co-authored with Prof. Wei Huang at Tongji University, was accepted by GIScience & Remote Sensing; honored to contribute as second author and big congratulations to Dr. Huan!
- 2025.10: 🔭 Joined Prof. Yu Liu and Prof. Fan Zhang’s team at Peking University as a remote research assistant.
- 2025.08: 🎉 Delivered an oral presentation at Hong Kong Polytechnic University after our paper was accepted to the Global Smart Cities Summit cum The 4th International Conference on Urban Informatics (GSCS & ICUI 2025).
- 2025.07: 🎉 My undergraduate thesis was accepted by 7th Asia Conference on Machine Learning and Computing (ACMLC 2025).
- 2025.06: 🎓 Graduated from Tongji University—grateful for the journey and excited to continue my studies at CMU.
- 2025.04: 🔭 Completed the SITP project under the supervision of Prof. Yujia Zhai in the College of Architecture and Urban Planning.
- 2025.01: 💼 Joined Shanghai Artificial Intelligence Laboratory as an AI Product Manager Intern.
- 2024.09: 🌏 Conducted research at ASTAR in Singapore under the supervision of Dr. Yicheng Zhang and Dr. Sheng Zhang.
- 2024.04: 🔭 Began my academic journey at Prof. Wei Huang’s lab in the College of Surveying and Geo-Informatics, Tongji University.
📖 Education
New York University
2026.09 – 2031.05 (expected)
Ph.D. student in Public Administration (Urban Policy)
Carnegie Mellon University
2025.08 – 2026.08
M.S. in Artificial Intelligence Systems Management
Tongji University
2021.09 – 2025.06
B.Mgt. in Information Management and Information System
💼 Experience
🔭 Research
- 2026.04 - Present, Research Assistant, Shanghai Key Laboratory of Urban Design and Urban Science (LOUD)
- 2025.10 - 2026.03, Research Assistant, Institute of Remote Sensing and GIS, Peking University, China
- 2024.09 - 2024.12, Research Officer Intern, A*STAR Institute for Infocomm Research, Singapore
- 2024.04 - 2025.04, Research Assistant, College of Architecture and Urban Planning, Tongji University, China
- 2024.04 - 2024.12, Research Assistant, College of Surveying and Geo-Informatics, Tongji University, China
💻 Industry
- 2025.01 - 2025.04, AI Product Manager, Shanghai Artificial Intelligence Laboratory, China.
- 2023.01 - 2023.02, Data Analyst, Shanghai Qiantan Emerging Industry Research Institute, China.
📝 Selected Papers
Peer-Reviewed
XXV ISPRS Congress
UrbanVGGT: Scalable Sidewalk Width Estimation from Street View Images
XXV ISPRS Congress, 2026.
Present UrbanVGGT, a measurement pipeline for estimating metrically scaled sidewalk width from a single street-view image using VGGT-based 3D reconstruction, semantic segmentation, and ground-plane fitting, achieving 0.25 m MAE on a Washington D.C. benchmark.
GSCS & ICUI 2025
Decoding Tourist Perception in Historic Urban Quarters with Multimodal Social Media Data: An AI-Based Framework and Evidence from Shanghai
Global Smart Cities Summit cum The 4th International Conference on Urban Informatics, 2025.
Developed an AI-powered multimodal framework to analyze tourist perception in historic Shanghai quarters, integrating image segmentation, color theme analysis, and sentiment mining for heritage-informed urban planning.
ACMLC 2025
Multimodal Deep Learning for Modeling Air Traffic Controllers Command Lifecycle and Workload Prediction in Terminal Airspace
Asia Conference on Machine Learning and Computing, 2025.
Designed a multimodal deep learning framework linking ATCO voice commands with aircraft trajectories to model workload dynamics, enabling intelligent command generation and scheduling support.
GIScience & Remote Sensing
A Spatiotemporal Adaptive Local Search Method for Tracking Congestion Propagation in Dynamic Networks
GIScience & Remote Sensing, 2025.
Proposed a spatiotemporal adaptive local search (STALS) method combining dynamic graph learning and spatial analytics to model and mitigate large-scale urban traffic congestion propagation.
Preprints
Preprint
CREG: Compass Relational Evidence for Interpreting Spatial Reasoning in Vision-Language Models
Preprint, 2026.
A training-free interpretability framework that maps vision-language model attributions into a polar coordinate system to reveal how VLMs reason about spatial relations.
Preprint
What We Lose, What We Gain: Spatio-temporal Patterns of Lost-and-Found Items in Qingdao Metro
Preprint, 2026.
Analyze 34,333 lost-and-found records across 173 metro stations to uncover spatio-temporal loss patterns, holiday effects, and station-level hotspots for transit service optimization.
Preprint
Do We Need a Robot Lane? A Simulation-Based Screening Framework for Sidewalk Delivery Robots
Preprint, 2026.
When should cities give sidewalk robots their own lane? An agent-based simulation framework that screens shared vs. dedicated space policies to answer this question.
🔬 Projects
RAGCache++: Cache-Aware Document Ordering for Low-Latency RAG Serving
Propose RAGCache++, a lightweight prompt-level optimization that reorders retrieved documents to maximize prefix sharing with cached sequences via a knowledge tree and greedy algorithm, reducing median TTFT by 20–33% with zero GPU memory cost and no serving-engine modification.
💬 Presentations
-
2026.07 - XXV ISPRS Congress 2026
UrbanVGGT: Scalable Sidewalk Width Estimation from Street View Images
Toronto, Canada -
2025.08 - Global Smart Cities Summit cum The 4th International Conference on Urban Informatics (GSCS & ICUI 2025)
A Multidimensional AI-powered Framework for Analyzing Tourist Perception in Historic Urban Quarters: A Case Study in Shanghai
Hong Kong Polytechnic University (PolyU), Hong Kong SAR, China -
2025.07 - 7th Asia Conference on Machine Learning and Computing (ACMLC 2025)
Multimodal Deep Learning for Modeling Air Traffic Controllers Command Lifecycle and Workload Prediction in Terminal Airspace
Hong Kong SAR, China
📫 Contact
- Email(CMU): kaizhent@cmu.edu
- Email(NYU): kt3275@nyu.edu
- Email(personal): wflps20140311@gmail.com
Please feel free to reach out if any of these research directions resonate with you. I'd be happy to chat!