Kang Liu Ann Arbor, MI | [email protected] www-personal.umich.edu/∼kangliu | www.linkedin.com/in/kang-liu
EDUCATION University of Michigan, Ann Arbor Ann Arbor, MI PhD, Control Systems, Electrical Engineering and Computer Science, GPA: 4.0/4.0 Sept. 2016 - Present Professional Development Chair of Tau Beta Pi Honor Society, Activity Chair of ECE Graduate Student Council Harbin Institute of Technology Harbin, CN Bachelor of Science, Mechanical Engineering, GPA: 3.91/4.00 Sept. 2012 - June 2016 Chiang Chen Overseas Fellowship (Hong Kong), Summer Research Grant by University of California, Los Angeles
WORK EXPERIENCE American Express Company (New York Office) July - Aug. 2020 Manager Intern, Machine Learning Research Group, Credit & Fraud Risk Unit · Used machine learning algorithms to detect the probability of customers being business owners and transactions with business purposes to improve credit line assignment. Model accuracy improved by 15% compared to the existing one. · Integrated the above metrics into existing production models, improved default prediction accuracy by 0.2%. · Researched on the impact of COVID-19 to US and international business owners, set project goals to incorporate effects of market shutdown. Developed strategies to clean-up datasets and recalibrate target variables for training. · Technical knowledge and skills involved included XGBoost, RNN, Python, SQL, cloud computing, Linux and git. PROJECT Thesis: The (α, β) Precision Theory for Production System Monitoring and Improvement Sept. 2017 - Present · Calculated the number of measurements required to obtain estimates of machine parameters, machine effciency and system throughput with a specified precision level, developed an upper bound on the data collection time. · First of its kind to address the optimal number of measurements problem in production operation research. · Delivered a presentation at the 2019 International Federation of Automatic Control Conference in Berlin, Germany. · Concluded the results in 1 peer reviewed conference paper and 2 journal papers. Chinese Ancient Painting Coloring Using Generative Neural Networks Jan. - Apr. 2019 · Implemented Deep Convolutional Generative Adversarial Network (DCGAN), and Wasserstein Generative Adversarial Network (WGAN) on MNIST and CIFAR-10 to color ancient Chinese painting sketches. · Compared the two implementations, DCGAN generates colored images with lower identification error rate.
A Vehicle Routing Problem (VRP) Variant Solved by Sequential Optimization Aug. - Sept. 2018 · Solved a VRP, with stochastic customer demand and chance constraints, using stochastic sequential optimization. · Delivered a presentation at the 2019 American Control Conference. Results published on IEEE conference proceedings.
PUBLICATION AS FIRST AUTHOR/LEADING AUTHOR
K. Liu, N. Li, I. Kolmanvosky, A. Girard, “A Vehicle Routing Problem with Dynamic Demands and Restricted Failures Solved Using Stochastic Predictive Control,” Proceedings of the IEEE 2019 American Control Conference, Philadelphia. [IEEE][ResearchGate] P. Alavian, Y. Eun, K. Liu (Leading), S. M. Meerkov and L. Zhang (Alphabetical Order), “The (α, β)-Precise Estimates of MTBF and MTTR: Definitions, Calculations, and Induced Effect on Machine Efficiency Evaluation,” Proceeding of the IFAC Manufacturing Modelling, Management and Control - 9th MIM 2019, Berlin, Germany, IFAC- PapersOnLine 52.13 (2019): 1004-1009. [IFAC][ResearchGate] P. Alavian, Y. Eun, K. Liu (Leading), S. M. Meerkov and L. Zhang (Alphabetical Order), “The (α, β)-Precise Estimates of MTBF and MTTR: Definition, Calculation, and Observation Time,” IEEE Transactions on Automation Science and Engineering. [IEEE][ResearchGate] P. Alavian, Y. Eun, K. Liu (Leading), S. M. Meerkov and L. Zhang (Alphabetical Order), “Precision of Machine Efficiency and System Throughput Estimates Based on (α, β)-Precise Estimates of MTBF and MTTR,” submitted to International Journal of Production Research.