Dr. Rui Song

Dr. Rui Song

Senior Principal Scientist

Amazon Inc.

About

I am currently a senior principal scientist at Amazon Inc. My current research interests include Machine Learning, Causal Inference, Precision Health, Knowledge Graph.

Download my CV . Contact me: songray@gmail.com

Interests
  • Machine Learning
  • Causal Inference
  • Precision Health
  • Knowledge Graph
Education
  • PhD in Statistics, 2006

    University of Wisconsin-Madison

  • B.S. in Statistics, 2001

    Peking University

Service

Please check out the Applied Reinforcement Learning online seminar series.

  • Associate Editor: Annals of Statistics 2022-
  • Associate Editor: Journal of the American Statistical
  • Association: Theory & method, 2021-
  • Guest Associate Editor: A special issue for Journal of Econometrics. 2021-2022.
  • Associate Editor: Journal of the American Statistical Association: Book review, 2018-
  • Associate Editor: Journal of Computational and Graphical Statistics, 2020-
  • Associate Editor: Biometrics, 2018-2020.

Grants

  • NSF grant DMS-2113637, 2021-2024, $200,000, “Offline Statistical Reinforcement Learning with Applications in Precision Health,” PI.
  • NSF grant DMS-1555244, 2016-2021, $400,000, “CAREER: Semiparametric and Machine Learning Approaches for Big Data Challenges in Precision Medicine,” PI.
  • NCSU Research and Innovation Seed Funding Program, 2014-2015, $15,000, “Machine Learning Methods for Annual Influenza Vaccine Update,” co-PI (PI, Dr. Osman Ozaltin from ISE department).
  • NCI grant P01 CA142538, 2012-2021, “Statistical Methods for Cancer Clinical Trials,” Co-I.
  • NSF grant DMS-1309465, 2013-2016, $109,999, “High-dimensional Multi-stage Statistical Learning with Application to Dynamic Treatment Regimens,” PI.
  • NCSU Faculty Research and Professional Development Award Grant, 2013-2014, $4,000, “Statistical Models, Methodologies and Related Theory For Developing Dynamic Treatment Regimens,” PI.
  • NSF grant DMS-1007698, 2010-2013, $100,000, “Variable Selection Methods in High Dimensional Feature Space,” PI.

Publications (by Year)

Preprints

2022

2021

2020

2019

2018

2017

2016

2015

2014

2012

2011

2010

2009

2008

2007

Software

  • Censored Rank Independence Screening

The R codes were developed for censored rank independence screening for high-dimensional right censored data. The R codes include all functions called can be found here. The original paper is appeared in Biometrika (2014).

  • Varying-coefficient Independence Screening for High-dimensional data

The R codes were developed for vary-coefficient independence screening for high-dimensional longitudinal data. The R codes include all functions called can be found here. The original paper is appeared in Statistica SInica (2015).

  • Penalized Q-learning for dynamic treatment regimes

The R codes were developed for statistical inference for penalized Q-learning in dynamic treatment regimes. The R codes include all functions called and a simulation setting can be found here. The original paper is appeared at Statistica Sinica (2014).

  • Subgroup detection and sample size calculation

The R codes were developed for testing the existence of a subgroup with enhanced treatment effect, and associated sample size calculation procedure for the subgroup detection test. An R package, named “subdetect” has been uploaded to CRAN. The codes were developed by Ailin Fan, Shannon Holloway, Wenbin Lu and Rui Song. The original paper is accepted for publication in Journal of American Statistical Association (2016).

  • Doubly robust estimation of optimal treatment regimes for maximizing t-year survival probabilities

The R codes were developed for implementing the doubly robust estimation methods proposed in the paper “On estimation of optimal for maximizing t-year survival probabilities” by Runchao Jiang, Wenbin Lu, Rui Song and Yong Marie Davidian (JRSSB, 2017, in press). The R codes (including a readme.txt file for a detailed description) and the AIDS data (ACTG175) used in the paper can be downloaded from here.

  • Concordance assisted learning (CAL) for estimating optimal individualized treatment regimes

The R codes were developed for implementing the CAL methods proposed in the paper “Concordance assisted learning for estimating optimal individualized treatment regimes” by Caiyun Fan, Wenbin Lu, Rui Song and Yong Zhou (JRSSB, 2017, in press). The R codes (including a readme.txt file for a detailed description) and the AIDS data (ACTG175) used in the paper can be downloaded from here.

  • Variable selection for optimal dynamic treatment regime

The R codes were developed for selecting important predictors in optimal dynamic treatment decision (i.e. those with qualitative interactions with treatments) based on two methods: the sequential advantage selection (SAS, Fan, Lu and Song, 2016, Annals of Applied Statistics) and high-dimensional A-learning (Shi, Fan, Song and Lu, 2018, Annals of Statistics, in press). An R package, named “ITRSelect” has been uploaded to CRAN.

  • Inference for quantile-adaptive dynamic treatment regime

The R codes were developed for estimation and inference for quantile-adaptive dynamic treatment regime An R package, named “quantoptr” has been uploaded to CRAN.

Students

Students I am currently (co-)advising/working with at NCSU:

  • Jianian Wang (expected to graduate in 2023)
  • Han Wang (expected to graduate in 2023, co-advised with Dr. Wenbin Lu)
  • Ye Shen (expected to graduate in 2024)
  • Lin Ge (expected to graduate in 2024)
  • Dale Gao (expected to graduate in 2024)
  • Richard Watson (expected to graduate in 2024)
  • Yang Xu (expected to graduate in 2025)
  • Haoyu Wei (expected to graduate in 2025)
  • Mohsen Sahraei Ardakani (Department of ECE)
  • Hunter Jiang

Former graduate students:

  • Runzhe Wan (graduated in 2022, first job at Amazon)
  • Hengrui Cai (graduated in 2022, co-advised with Dr. Wenbin Lu, first job is assistant professor at UC Irvine)
  • Haoyu Chen (co-advised with Dr. Wenbin Lu, graduated in June - 2021, first job at Pinterest)
  • Sheng Zhang (graduated in June 2021, first job at Amazon)
  • Kevin Gunn (co-advised with Dr. Wenbin Lu, graduated in Jul 2020, first job at Liberty Mutual)
  • Ye Liu (graduated in June 2020, first job at Google)
  • Chaowen Zheng (co-advised with Dr. Yichao Wu, graduated in March 2020, first job at Bank of America)
  • Liangyu Zhu, PhD (co-advised with Dr. Wenbin Lu, graduated in Dec 2019, first job at Google)
  • Chengchun Shi, PhD (co-advised with Dr. Wenbin Lu, graduated in Jun 2019, first job is assistant professor at LSE)
  • Shuhan Liang, PhD (co-advised with Dr. Wenbin Lu, graduated in May 2018, first job at Google)
  • Ailin Fan, PhD (co-advised with Dr. Wenbin Lu, graduated in May 2016, first job at Chase)
  • Shikai Luo, PhD (co-advised with Dr. Subhasis Ghoshal, graduated at May 2016, first job at Quantlab)
  • Runchao Jiang, PhD (co-advised with Dr. Wenbin Lu, graduated in May 2015, first job at Facebook)
  • Neal Jorgensen, MS (graduated in December 2010 at Colorado State University)

Visiting undergraduate students:

  • Yunan Gao, BS (Tsinghua University, May 2017-August 2017)
  • Cheng Ma, BS (Tsinghua University, May 2017-August 2017)
  • Lantian Xu, BS (University of Science and Technology of China, May 2017-August 2017)

Teaching

At North Carolina State University:

  • ST 361 Introduction to Statistics for Engineers
  • ST 508 Statistics For the Behavioral Sciences II
  • ST 511 Introduction to Statistics for Biological Sciences
  • ST 563 Introduction to Statistical Learning
  • ST 745 Analysis of Survival Data
  • ST 790 Financial Statistics
  • ST 810 Topics in High-dimensional Statistical Inference

At Corolado State University:

  • ST 305 Sampling Techniques
  • ST 501 Statistical Science
  • ST 640 Design and Linear Modeling
  • ST 740 Introduction to Empirical Processes and Semiparametric Inference

Contact