Personal Update:
I am currently involved in the Mens, Manus, and Machina (M3S) project on Artificial Intelligence (AI) led by Prof. Jinhua Zhao at MIT. M3S applies AI and machine learning practically, addressing technology design, human skill development, and societal adaptation to AI, automation, and robotics, with the goal of creating inclusive and resilient solutions with global impact.

About Me


I am Xinyu Chen (陈新宇), now a Postdoctoral Associate at MIT (or visit MIT sites), leading the spatiotemporal data modeling project on GitHub. I am currently working on developing some theoretical machine learning methods (e.g., matrix/tensor decomposition) for modeling a wide range of spatiotemporal data. These data are by nature multidimensional tensors collected from real-world systems, including human mobility, trajectory data, traffic flow, fluid flow, climate/weather data, and energy consumption data. The specific scientific problems are 1) estimating and forecasting missing values of spatiotemporal data in real-world scenarios, 2) discovering dynamic patterns from spatiotemporal data (e.g., human mobility and climate data), and 3) generating high-resolution output from the low-resolution input data according to the spatiotemporal domain knowledge (e.g., speed field reconstruction and high-resolution climate variables generation). My Ph.D. thesis is entitled as "Matrix and Tensor Models for Spatiotemporal Traffic Data Imputation and Forecasting".

I like mathematics and computer science related stuff. My research interests include, but are not limited to, machine learning, spatiotemporal data modeling, intelligent transportation systems, urban science, and AI for science. My research philosophy comes from an ancient Chinese philosopher and writer Laozi, who stated "大道至简" (make it as simple and clear as possible). I am a strong advocate of open-source and reproducible research. To contribute more ideas to the research community, I am trying to make the datasets and Python codes of our research publicly available on GitHub. I am now leading some innovative open-source projects on GitHub (with 500+ followers), and they have accumulated more than 4.5k stars. Two most popular GitHub repositories are transdim (providing machine learning algorithms for transportation data imputation and prediction, 1.2k+ stars) and aswesome-latex-drawing (providing academic drawing examples in LaTeX, 1.3k+ stars).

Fortunately, I have received several awards during my PhD study, including IVADO PhD Excellence Scholarship ($100k by Institute for Data Valorisation (IVADO)) and CIRRELT PhD Excellence Scholarship ($5k by Interuniversity Research Centre on Enterprise Networks, Logistics and Transportation (CIRRELT)). I would like to thank IVADO and CIRRELT for funding my Ph.D. research.

Selected News

(For the full list of news please check the News tab.)

Archive

vistors & views since September 2021

drawing
drawing