Hi, my name is

Yutong Xia

PhD Student @ NUS

Hi! I am Xia Yutong (夏宇彤), currently a Research Fellow at School of Computing (SoC), National University of Singapore (NUS), working with Prof. Roger Zimmermann. I received my Ph.D. in Data Science from NUS, under the supervision of Prof. Roger Zimmermann. My research interests lie in spatio-temporal data mining, urban computing, and time series analysis. My recent work focuses on leveraging causal AI techniques to better understand and model complex spatio-temporal systems.

I am open to research collaborations in causality AI for time series, spatio-temporal and urban data. Please contact me at: yutong.x@outlook.com.

News

[Apr. 2026] I successfully defended my PhD thesis! Special thanks to my supervisor Prof. Roger for all the support throughout my PhD journey! :)
[Jun. 2025] We will give a tutorial on Foundation Models for Spatio-Temporal Data at KDD’25. Check out our survey paper here.
[Apr. 2025] Invited talk about Towards Predictive Cities: Learning Spatio-Temporal Graphs Data in the AI Era at City+Talk.
[Dec. 2024] Two papers about graph data augmentation and air quality inferrence were accepted by AAAI'25.
[Nov. 2024] Calling for papers! We are organizing WebST Workshop 2025 - Spatio-Temporal Data Mining from the Web at WWW'25. We are calling for submissions (4-8 pages).
[Nov. 2024] One paper about ST forecasting was accepted by KDD'25 ADS Track.
[May. 2024] One paper about large-scale delivery dataset was accepted by KDD'24 ADS track.
[May. 2024] Two papers about Singapore parking dataset and spatio-temporal field neural network were accepted by IJCAI'24 AI for Social Good track.
[Jan. 2024] I was honored to receive the NUS SoC Research Achievement Award. Thanks Prof. Roger for the supports!!
[Sept. 2023] One paper about spatio-temporal causal inference was accepted by NeurIPS'23.
[Sept. 2023] One large-scale traffic benchmark was accepted by NeurIPS'23 DB Track.
[Sept. 2023] I was honored to receive the NUSGS Research Incentive Award!
[Sept. 2023] One paper about spatio-temporal diffusion model was accepted by SIGSPATIAL'23.
[Jan. 2023] I will attend AAAI'23 at Washington DC in person ✈️. (It is my first time attending a conference!! XD)
[Nov. 2022] One paper about large-scale air quality prediction using Transformers was accepted by AAAI'23 as an oral presentation.
[Aug. 2022] One paper about travel mode choice prediction was accepted by Travel Behaviour and Society.
[Aug. 2022] One paper about accessibility equality was accepted by Sustainable Cities and Society.
[Jan. 2022] I started my Ph.D. candidature at National University of Singapore.
[Dec. 2021] I graduated from University College London and obtained my master’s degree with a distinction.

Research Vision

The city we observe, the city we model, and the city we live in are not the same. My research uses causal and spatio-temporal methods to understand when these worlds align—and when they do not—by studying hidden mechanisms, alternative futures, and generalization across cities and changing environments.

The city we live in, the city we observe and the city we model, connected in a cycle by observation and selection, learning and inference, and decision and intervention, asking when these worlds align Observation & Selection Measurement · Sampling · Missingness Decision & Intervention Deployment · Response · Adaptation Learning & Inference Discovery · Prediction · Generation When Do These Worlds Align? The City We Live In The CityWe Live In People · Infrastructure Institutions · Environment The City We Observe The CityWe Observe Sensors · Platforms Administrative Records Partial · Selective · Confounded The City We Model The CityWe Model Inferred Mechanisms · Predictions Alternative Futures
Core questions
Hidden mechanisms
What can be recovered from partial and confounded observations?
Selected work → CaST · CaPulse · UrbanCIA
Alternative futures
How can models represent interventions and counterfactual trajectories?
Selected work → CaTSG
Unseen cities
What transfers across cities and changing environments?
Selected work → RE-BNN · STFM Survey
Methods
Spatio-Temporal Data MiningCausal Inference & DiscoveryGenerative ModelingFoundation Models

Community Services

  • Organizer of WWW'25'26 Workshop: WebST (Publicity Chair)
  • Contributor of KDD’25 Tutorial: Foundation models for spatio-temporal data science
  • Session Chair: KDD’25
  • Senior PC Member: CIKM’25’26
  • PC Member & Reviewer: ICML’25’26, ICLR’25’26’27, NeurIPS’24’25’26, KDD’25’26’27, AAAI’26’27, AISTATS’25’26, WWW’26, CIKM’26, SIGSPATIAL’25, ECML’23, UrbanComp’25
  • Journal Reviewer: TPAMI, TKDE, TDSC, Neurocomputing, TNNLS, Pattern Recognition, TMM, IJGIS, Scientific Reports, Ann. GIS, TMLR

Professional Experience

Research Fellow - National University of Singapore
Mar 2026 - Present
I currently work as a Research Fellow at School of Computing, National University of Singapore (NUS) in Singapore, working with Prof. Roger Zimmermann. I work on IBM benchmarking and geographical foundation models.
I worked as a Research Intern at Microsoft Research Asia in Beijing, China, mentored by Dr. Jiang Bian and Dr. Chang Xu. I worked on a project about causal time series generation via diffusion models.
I visited at JTL Urban Mobility Lab at Massachusetts Institute of Technology (MIT), supervised by Prof. Jinhua Zhao. I worked on a project about scaling urban causal inference with large language models, in collaboration with Ao Qu.
I worked as a Research Intern at Alibaba Cloud, Alibaba Group in Hangzhou, China, mentored by Dr. Lunting Fan and Dr. Yingying Zhang. I worked on a project about time series anomaly detection.