Xie's Research Group

Our group's current research focus is on causal machine learning and causal AI.
We develop methods to address a major challenge in causal inference from
observational data, the absence of perfect interventions. We explore how
causal inference, machine learning, and AI can enhance one another.
We also study causal generative AI methods for treatment effectiveness
evaluation and precision medicine.


Our team won the Covid Causal Diagram DREAM Challenge.

A fundamental question: When is causal inference possible?
Bibliography (internal site for group members)

Causal machine learning (ML) example from Netflix, credit to Netflix TechBlog



Causal machine learning is related to the following research areas,
which are often referred to with different names.

  1. Causal feature representation learning
  2. Causal graph learning
  3. Out-of-distribution generalization
  4. Treatment effect estimation, heterogenous treatment effects
  5. Counterfactual estimation
  6. Real World Data (RWD) or Real World Evidence (RWE)



Grant support





Past Publications

  1. Cauchy combination test, published in JASA (Journal of the American Statistical Association)
    A test for weak and sparse alternatives with analytic p-value calculation and under arbitrary dependency structures
  2. Another paper in JASA
    "Accurate and Efficient P-value Calculation via Gaussian Approximation: a Novel Monte-Carlo Method"
  3. In the Annals of Applied Statistics 2018, Vol. 12, No. 1, 567-585.
    "Powerful test based on conditional effects for genome-wide screening"

Google Scholar Citation

Curriculum Vitae

Education

1994: B.S. in Statistics
Department of Probability and Statistics, Peking University

1997: M.S. in Statistics
Department of Probability and Statistics, Peking University

2000: Ph.D. in Statistics
Department of Statistics, University of California at Los Angeles


Research Group


Past group members in recent years