BIOGRAPHY

I am a PhD student at UC San Diego department of Computer Science and Engineering, advised by Prof. Rose Yu and Prof. Yian Ma. I received my Bachelor degree of science in Applied Math, Physics, and Computer Sciences from the University of Wisconsin-Madison in 2020. My research interests are in Bayesian Deep Learning, Spatiotemporal Forecasting, and AI for Health.

Publications

  • Accelerating Stochastic Simulation with Interactive Neural Processes
  • Dongxia Wu, Matteo Chinazzi, Alessandro Vespignani, Yian Ma, Rose Yu.
    Under review
  • Quantifying Uncertainty in Deep Spatiotemporal Forecasting
  • Dongxia Wu, Liyao Gao, Xinyue Xiong, Matteo Chinazzi, Alessandro Vespignani, Yian Ma, Rose Yu.
    ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) 2021
  • DeepGLEAM: A hybrid mechanistic and deep learning model for COVID-19 forecasting
  • Dongxia Wu, Liyao Gao, Xinyue Xiong, Matteo Chinazzi, Alessandro Vespignani, Yian Ma, Rose Yu.
    Under review
  • A deep learning based automatic defect analysis framework for In-situ TEM ion irradiations
  • Mingren Shen, Guanzhao Li, Dongxia Wu, et al.
    Computational Materials Science, 2021, 197: 110560.
  • Multi defect detection and analysis of electron microscopy images with deep learning
  • Mingren Shen, Guanzhao Li, Dongxia Wu, et al.
    Computational Materials Science, 2021, 197: 110560.

    Experiences

    Research Assistant

    2020 - Present
    UC San Diego, La Jolla, CA
    • Spearheaded the deep learning uncertainty quantification study for spatiotemporal forecasting, compared Bayesian and Frequentist methods on traffic and COVID-19 predictions.
    • Created the DeepGLEAM model, which has been included in the CDC COVID-19 forecasting ensemble.
    • Design a neural process model for COVID-19 predictions to enable counterfactual reasoning.

    Research Assistant, NextGen Fellow

    2018 - 2020
    UW-Madison, Madison, WI; Citrine Informatics Inc., Redwood City, CA

    Dislocation Loop Analysis on TEM Video with Deep Learning Methods

    • Created an automatic and superfast dislocation loop detection framework for TEM Video of FeCrAl alloy under in-situ ion irradiation based on the YOLOv3 network.
    • Lead static property analysis for dislocation loops’ geometric information.

    Awards/Honors

    HDSI Ph.D. Fellowship (2021-2024), top 10 in admitted HDSI Ph.D. students - UC San Diego (2021)
    College of Letters & Science Dean’s List - UW–Madison (2018, 2019)
    Nextgen Fellowship - Citrine Informatics Inc. (2018)
    REU Program Fellowship - NSF (2018)
    Undergraduate Summer Scholarship - UW-Madison (2018)
    The First Prize in China Region and the Fourth Prize in the final - The Third Hong Kong International Chamber Music Competition (2016)

    Skills & Proficiency

    Python & Pytorch

    Tensor Flow & Keras & Matlab & LaTeX

    Java & Mathematica

    C & C++ & C#