Graduate Research Assistant

Leo (Liang) Yang

*********@*****.*** – 650-***-****

EDUCATION

Stanford University Stanford, CA

PhD candidate, Geological Sciences (Geostatistics), GPA: 3.82 Expected Graduation: Mar. ’21 Research: Spatial-temporal statistics; machine learning; data analysis; risk assessment; energy prediction Relevant coursework: Statistical modelling; Machine learning; Deep learning; Data mining; Geostatistics Sun Yat-sen University Guangzhou, China

MS and BS, Geological Engineering and Geoinformatics Jun. ’16 and Jun. ’13 Research: Spatial-temporal modelling and prediction; data analysis; optimization Relevant coursework: Database and SQL; C++ programming; Data structures and algorithms PROJECTS

Spatial Prediction with Machine Learning Methods and Risk Assessment Sept. ’16 – Present

(Research) Developed a bagging-like machine learning method for combining multiple predictions of spatial models into a single prediction to enhance the spatial model’s prediction accuracy and quality. (Python with scikit-learn)

(Research) Adapted a Gaussian process method (co-kriging) for spatial prediction and stochastic simulation, with parallel computation. (Python)

(Research) Created a Markov chain based algorithm for visualizing uncertainty of ore bodies using level sets (a computer vision method) with animated stochastic motion. (C++) Rainforest Connection Species Audio Detection with Convolutional Neural Network (CNN) Dec. ’20 – Feb. ’21

(Kaggle competition) Implemented a CNN model for detecting bird and frog species in a tropical soundscape. (Python) Exploratory Data Analysis of Large-scale Complex Industrial Geospatial Data Sept. ’18 – Sept. ’20

(Research) Designed a framework that combines statistical methods, such as hypothesis testing and unsupervised learning, for large-scale data analysis to improve accuracy in energy forecasting. (Python, R and SQL) Automatic Fashion Generation with Attentional Generative Adversarial Network Sept. ’19 – Dec. ’19

(Course project) Implemented an attentional mechanism for conditionally generating images (encoded using CNN) from texts (encoded using Bi-LSTM), using parallel computation, deployed on Azure. (Python) Music Transcription Using Deep Learning Sept. ’17 – Dec. ’17

(Course project) Compared and evaluated LSTM and DNN for music transcription tasks with hyperparameter tuning, using parallel computation, deployed on AWS. (Python) Stochastic Simulation of Subsurface with Multiple-point Geostatistics Jul. ’12 – Jun. ’16

(Research) Wrote a new EM-like optimization algorithm for subsurface prediction and simulation. The algorithm learned from training images with improved reproducibility of spatial patterns for stochastic simulation. (C++) Analysis of Multi-source Geospatial Data Jul. ’12 – Jun. ’16

(Research) Analyzed data with uncertainties and implemented an entropy based data aggregation method. (SQL and C#) EXPERIENCE

Graduate Research Assistant – Stanford Center for Earth Resources Forecasting Sept. ’16 – Present Role: Developed statistical algorithms to analyze large-scale complex spatial data, perform predictions and assess risks. Data Analyst and Research Assistant – Sun Yat-sen University Jul. ’12 – Jun. ’16 Role: Designed and implemented methods for spatial data analysis, subsurface modelling and risk assessment. SKILLS

Programming: Python (pandas, numpy, scikit-learn, jupyter notebooks, etc.), R, SQL, Matlab, C, C++, C#, Java, swiftUI Deep learning framework and cloud infrastructure: Pytorch, Tensorflow, AWS, AZure, Google Cloud Github: https://github.com/liangy396

SELECT PUBLICATIONS

Yang, L., Hyde, D., Grujic, O., Scheidt, C. and Caers, J., 2019. Assessing and visualizing uncertainty of 3D geological surfaces using level sets with stochastic motion. Computers & Geosciences, 122, pp.54-67. Yang, L., Hou, W., Cui, C. and Cui, J., 2016. GOSIM: a multi-scale iterative multiple-point statistics algorithm with global optimization. Computers & Geosciences, 89, pp.57-70.

Contact this candidate