Daniel Iong

Daniel Iong

PhD Candidate in Statistics

University of Michigan, Ann Arbor

Education
  • PhD Statistics 2017 - 2023

    University of Michigan, Ann Arbor

  • BS Statistics, BA Economics 2013 - 2017

    University of California, Davis

Interests
  • Probabilistic/Bayesian modeling
  • Time series forecasting
  • (Statistical) machine learning
  • Applied Statistics

Experience

 
 
 
 
 
Orbital Insight
Data Scientist Intern
May 2022 – Aug 2022 Palo Alto, CA
  • Developed algorithm based on convolutional neural networks and other computer vision methods to detect rare GNSS interference events in geolocation (AIS, ADS-B) data as a key deliverable for a Department of Defense contract.
  • Trained convolutional neural networks on generated synthetic data using PyTorch on AWS EC2 instances.
 
 
 
 
 
NASA Goddard Space Flight Center
Research Intern
Jun 2021 – Aug 2021 Virtual
  • Mentors: Charles N. Arge, Michael Kirk, Daniel Da Silva
  • Identified novel application of dynamic time warping for model evaluation of a physics-based solar wind model
  • Created web app using Dash and Plotly packages in Python to visualize dynamic time warping (see below).

Projects

.js-id-completed
Variational inference for robust Gaussian process regression
We developed a scalable approximate inference algorithm for robust Gaussian process regression with contaminated normal noise. Initial experiments with simulated and actual data are promising.
Variational inference for robust Gaussian process regression
Explainable ML for space weather forecasting
We trained gradient boosted trees to forecast the SYM-H index several hours ahead using solar wind data and used SHAP values to explain predictions.
Modeling heterogeneous causal mechanisms in epidemiology with observational data
We introduced the notion of ‘mechanistic heterogeneity’ in Mendelian Randomization, developed a latent mixture method to model it, and used it to estimate heterogeneous causal effects between HDL cholesterol and coronary heart disease.

Publications

(2022). Global Sensitivity Analysis and Uncertainty Quantification for Background Solar Wind using the Alfvén Wave Solar Atmosphere Model. Space Weather.

DOI PDF

(2022). New Findings From Explainable SYM-H Forecasting Using Gradient Boosting Machines. Space Weather.

Project DOI PDF

(2020). A Latent Mixture Model for Heterogeneous Causal Mechanisms in Mendelian Randomization. Submitted to Annals of Applied Statistics.

Project arXiv PDF