WENBO SUN 2013 Medford Rd, Apt 161, Ann Arbor, 48104, MI, USA (1)-734-834-4191 • [email protected]

EDUCATION: UNIVERSITY OF MICHIGAN, 4TH YEAR PHD CANDIDATE (Expected) Dec 2018 Department of Industrial and Operation Engineering Overall GPA: 4.0/4/0 Related Coursework: Statistical Quality Control, Dynamic Programming, Machine Learning , Graphical Model , Spatial Statistics, Dynamic Treatment Regime UNIVERSITY OF MICHIGAN, APPLIED MASTER OF STATISTICS Department of Statistics

Dec 2013

PEKING UNIVERSITY, CHINA, B.S. IN PROBABILITY & STATISTICS July 2012 School of Mathematical Science Honors: • The International Mathematical Contest in Modeling, Honorable Mention, 2010 and 2011 • The National Mathematical Contest in Modeling, Second Prize, 2009 and 2010

RESEARCH EXPERIENCE: Research Interests Prof. Judy Jin & Matthew Plumlee, Department of IOE June 2016-Present • Design Optimization Based on Inexact Computer Models: Calibrate inexact computer model based on physical experiment data to search for optimal solutions. The proposed algorithm results in a consistent and robust estimator. Graduate Research Assistant Prof. Judy Jin & Shan Bao, Department of IOE & UMTRI June 2016-Present • Online Detection Algorithm Development for Intelligence Vehicle: Build statistical model to detect distracted driving behaviors based on real-time vehicle motion signals. The resultant model can guide autonomous vehicles to follow normal driving vehicles. Both the misdetection rate and the false alarm rate are below 15%. Graduate Research Assistant Prof. Judy Jin & Jonathan Rupp, Department of IOE & UMTRI Jan 2014-June 2016 • PCA Functional Data Analysis: Estimate confidence bounds for kinematic response signals during crash to evaluate vehicle safety. A totally new method was developed based on PCA and Gaussian Process Regression in the field of biomechanics. Graduate Research Assistant Prof. Felipe Csaszar, Ross School of Business Sept 2012-Feb 2013 • R-to-C integrated programming: Rewrote the R program by Rcpp package to accelerate the loops; The improvement of speed in loops is offset by time spent in calling C from R. • C-to-F integrated programming: Re-organized the R program in C-to-F interface by using LAPACKE package for Q-R factorization; The C-version program is 40 times faster than the R-program. Undergraduate Research Assistant Prof. Zhi Geng, School of Mathematical Science Oct 2009–Jun 2012 • Bayesian Networks and Causal Inference: Successfully implemented five algorithms in C using data generated from the Alarm Network, including IC, PC, IAMB, PCMB, and BFMB algorithms • Gene Regulatory Networks: Improved an algorithm to enhance robustness in building the minimal network for the budding-yeast and the fission-yeast models by adding random noise to the transmission delay.

WORK EXPERIENCE: Computing Consultant Internship, ICPSR, University of Michigan Jun 2013-Aug 2013 • Provided support for statistical software, including EVIEWS, Excel, SPSS, R, SAS and STATA assist with the user’s computing and statistical problem

Statistical Modeling Internship, QS, P&G, China May 2013-Jun 2013 • High Dimensional Data Project: Investigated key variables for experiment by self-designed variable selection process combining nonlinear support vector regression with stepwise selection process; The set of selected variables was 1000 times smaller than the original set and were validated by the chemical experimenters. • Methodology Implementation: Reviewed some academic papers for better applying Lasso regression with higher stability across different data sets • R-to-Jmp programming: Called R program in Jmp to integrate the whole algorithm in one simple software, skipping the process of copying data between different software and running complicated R code from Rscript. As a result, experimenter without statistical background can apply the program by simply importing data to Jmp and getting fully analyzed result as well as controllable diagrams. IOE student representative, Michigan Student Symposium for Interdisciplinary Statistical Sciences Sept 2015-March 2016. • Organized the local academic conference in the statistic area together with other committee members as the IOE student representative. Over 200 abstracts was submitted to the conference. Invited and accomodated Prof. Michael Jordan as the keynote speaker of the year. Co-founder, Two-dimensional Manifold Studio Sept 2010-Jun 2012 • Founded and ran the studio focusing on details to create artwork by exploring the visualization of mathematics; Designed more than 10 T-shirts with mathematical meaning during the 4 years; Recorded and edited 4 promotional videos for school of mathematical sciences; Fundraised and supported two disabled students from low-income families each year Co-founder, Sweet Foundation Jan 2012-Present • Founded and ran the social venture dedicated to helping people who are suffering from hunger. The fundraising process is based on selling baked goods made by students from University of Michigan. During the past 4 years, 700 dollars were raised and sent to African organizations to help hungry people there on average every year.

COMPUTER SKILLS: • • • •

Operating Systems: Windows, UNIX Programming Languages: Proficiency with C, Pascal, FORTRAN, Latex, Latex-Beamer Mathematical Languages: R, Python, SAS, Stata, Mathematica, Matlab, SPSS, Lingo, 1stOpt, WinBugs Interface Programming: C-to-F, R-to-C, C-to-Matlab, R-to-Mathematica, R-to-Jmp

PUBLICATIONS

• Zhang & Sun, et al., 2012, Robust Motif In Regulatory Network, Journal of Theoretical Biology. (submitted) • Sun, et al., 2016, A Method for Developing Biomechanical Response Corridors Based On Principal Component Analysis, Journal of Biomechanics. (accepted) • Feng, et al., 2017, Estimating Lead Vehicle Kinematics Using Host Vehicle Camera-Based Sensory Data. (accepted) • Sun, et al., 2017, Online Detection for Distracted Driving Behaviors (unpublished manuscript) • Sun, et al., 2017, Robust Design Using an Inexact Simulation Model with Physical Experiment Data (unpublished manuscript) • Sun, et al., 2017, Develop Confidence Bands for Multiple Dimensional Functional Responses. (in progress)

PRESENTATIONS AND POSTER SESSIONS

• Poster session: Sun, et al., A New Statistical Method for Developing Biomechanical Response Corridors • Contributed presentation: Sun, et al., Develop Confidence Bands for Multiple Dimensional Functional Responses, INFORMS 2016 • Invited presentation: Sun, et al., Functional Data Based Computer Model Calibration for Design Optimization, INFORMS 2017 • Invited presentation: Sun, et al., Robust Design Using an Inexact Simulation Model with Physical Experiment Data, MSSISS 2018