I am Haotian Xiang (向昊天). The pronunciation is close to /haʊˈtiːən/ /ʃjɑŋ/.

I am currently a senior master student in the Electrical Engineering department of Columbia University. Before that, I have completed my B.Sc. with first class honor in the Electrical Information Engineering department from University of Electronic Science and Technology of China (UESTC).

My research interests include Generative Models, Intelligent Transportation Systems, and Active Learning. I am familiar with Urban Spatial-Temporal Prediction, particularly in the areas of Human Mobility Forecasting and Traffic Flow Forecasting. I am passionate about training more Robust and more General models to survive from Out-of-Distribution (OOD) data in Real-World Scenarios.

Long-Live AI4Science !

I am the Transformer Generation !

As a Transformer Boomer, my research experience has been closely intertwined with the Transformer based models. From the inception of the Transformer Encoder to Visual Transformer (ViT), and further into the realms of Multimodal domain (ALBEF) and Spatial-Temporal Attention (ASTGCN), my work has revolved around exploring the applications of the Transformer model in various domains. I’d like to explore more possibilities of applying Transformer based model to tackle frontier issues.

Activities

• [Sep 13, 2023] Our Active Learning Pipeline for Animal Behaviour Videos was completed and aiming for Cosyne 2024, working with mates in Paninski’s lab.
• [Aug 15, 2023] Our draft paper “diffIRM: A Diffusion-Augmented Invariant Risk Minimization Framework for Spatiotemporal Prediction over Graphs” for AIStat 2024 (see abstract) was finished, working with mates in Ditect lab.
• [Jun 01, 2023] I joined Liam Paninski’s Lab in Zuckerman Institute as a graduate researcher, focusing on Active Learning in analysis of Animal Behaviour Videos.
• [May 25, 2023] Our paper “Cross- and Context-Aware Attention Based Spatial-Temporal Graph Convolutional Networks for Human Mobility Prediction” was submitted to ACM Transactions on Spatial Algorithms and Systems workshops 2023, working with mates in Ditect lab.
• [Feb 04, 2023] I joined Professor Xuan Di’s Ditect lab, Data Science Center, Smart City Section, as a graduate researcher, focusing on Intelligent Transportation Systems, especially for Human Mobility Forecasting and Traffic Flow Forcasting.
• [Jan 08, 2023] I built a Stable Diffusion Model to generate X-ray Diffusion Spectrum Experiment data in Chemical Filed and build an ALBEF liked multi-modal model to perform Classification Task, supervised by Professor Tong Geng in the Rochester University.