Education

• M.S. in Data Science, University of Virginia (UVA), GPA 3.94/4.00, 2019

• Ph.D. in Chemistry, University of Science and Technology of China, 2013

  • Visiting Doctoral Program, Pennsylvania State University

• B.Sc. in Chemical Engineering, HeFei University of Technology, 2008

Research Experiences

Surface-Wave Waveform Classification Using Machine-learning Algorithms

• Collaborated with researchers from Oak Ridge National Laboratory and other institutions to automatically classify seismic observations using a labeled dataset of around 400,000 records

• Developed logistic regression, K-nearest neighbors, support vector machine, and artificial neural network algorithms; compared with random forests (developed by a collaborator)

• Artificial neural network and random forests outperform other algorithms and achieved an accuracy of over 92%. A manuscript has been submitted to Seismological Research Letters and currently in revision.

Prediction of Patients Deterioration in the Cardiac Wards

• Built logistic regression, random forests, extreme gradient boosting and super learner models using two million records to predict if a patient needs to be transferred to the intensive care unit (ICU) in 24 hours

• Applied Synthetic Minority Oversampling Technique (SMOTE) to handle class imbalance; built super learner by “stacking” the three algorithms together; explored machine-learned features using Autoencoder algorithm; compared four machine learning algorithms using different metrics

• Super learner leads the scoreboard with an area under curve (AUC) of 0.79 and an F1 score of 0.24. The paper has been accepted and presented in 2019 Systems And Information Engineering Design Symposium.

Aerial Cactus Detection Using Convolutional Neural Network (CNN)

• Built a CNN model with TensorFlow using a dataset of 17,500 images to predict whether a low-resolution photo contains a cactus

• Investigated the effect of padding, optimizers, epochs and number of convolutional layers using loss/accuracy of training and validation sets

• The optimized architecture achieved an accuracy of 99% on the test data set.

Stock Price Movement Prediction Using Deep Learning Methods

• Built random forests, Feed Forward Neural Network (FFNN), CNN and Recurrent Neural Network (RNN) models with TensorFlow to predict trading price movement (up, down, or stationary) of the Netflix stock

• Computed six stationary features to mitigate autocorrelation and reduce noise

• The best performance was obtained by the RNN model with a Cohen’s Kappa coefficient of 0.53.

Multi-class Cuisine Detection Using Recurrent Neural Network (RNN)

• Built a RNN model with TensorFlow using a JSON-formatted dataset to detect a cuisine type (20 classes) based on the recipe ingredients (varying lengths)

• Used the pre-trained word embedding GloVe to represent each word as a numeric vector; investigated the effect of cell types including basic cell, long short-term memory (LSTM), and gated recurrent unit (GRU); tested number of layers and dropout optimization

• The best model (used the GRU cell) resulted in a Cohen’s Kappa coefficient of 0.72.

Data Visualization for Features Comparison

• Built a HTML file using the JavaScript library D3.js to dynamically show and compare the features and their first two principle components between two groups of records

• Users can view data easily and in meaningful ways using interactive interfaces such as clicking points, dragging slider, changing data representation. The visualization helps on feature engineering and selection.

Skills

Language & Software

• Python (including scikit-learn, pandas, numpy, Matplotlib, seaborn, TensorFlow, BeautifulSoup)

• R (including ggplot2, tidyr, dplyr, plotly),

• SAS

• SQL

• MATLAB

• Spark

• Tableau

• Adobe Illustrator

• HTML

• JavaScript

• Rawgraphs

• AutoCAD

• Photoshop

Machine Learning Skills

• Logistic Regression

• Polynomial Regression

• Random Forests

• Support Vector Machine

• Feed Forward Neural Network

• Convolutional Neural Network

• Recurrent Neural Network

• Autoencoder

• K-nearest Neighbors

• K-means Clustering

• Principle Component Analysis (PCA)

Selected Publications

1. Chengping Chai, Jonas A Kintner, Kenneth M. Cleveland, Jingyi Luo, Monica Maceira, Charles Ammon, Automatic waveform quality control for surface waves using machine learning, Seismological Research Letters, (2022), https://doi.org/10.1785/0220210302.

2. Justin Niestroy, Jiangxue Han, Jingyi Luo, Runhao Zhao, Douglas E. Lake, Abigail Flower (authors contributed equally), Prediction of Decompensation in Patients in the Cardiac Ward. 2019 Systems and Information Engineering Design Symposium (SIEDS), Charlottesville, VA, USA (2019), https://doi.org/10.1109/SIEDS.2019.8735602.

3. Jingyi Luo, Cuiming Wu, Yonghui Wu, TongwenXu. Diffusion dialysis of hydrochloric acids with their salts: Effect of co-existence metal ions. Separation and Purification Technology, 118 (2013) 716-722, https://doi.org/10.1016/j.seppur.2013.08.014.

4. Yonghui Wu, Jingyi Luo, Liliang Zhao, Gencheng Zhang, Cuiming Wu, TongwenXu, QPPO/PVA anion exchange hybrid membranes from double crosslinking agents for acid recovery. Journal of Membrane Scienc, 428 (2013) 95-103, https://doi.org/10.1016/j.memsci.2012.10.018.

5. Yonghui Wu, Jingyi Luo, Lulu Yao, Cuiming Wu, Fulin Mao, Tongwen, Xu, PVA/SiO2 anion exchange hybrid membranes from multisilicon copolymers with two types of molecular weights. Journal of Membrane Science, 399-400 (2012) 16-27, https://doi.org/10.1016/j.memsci.2012.01.019.

6. Jingyi Luo, Cuiming Wu, TongwenXu, Yonghui Wu, Diffusion dialysis - concept, principle and applications. Journal of Membrane Science, 366 (2011) 1-16, https://doi.org/10.1016/j.memsci.2010.10.028.

7. Jingyi Luo, Cuiming Wu, Yonghui Wu, TongwenXu, Diffusion dialysis processes of inorganic acids and their salts: the permeability of different acidic anions. Separation and Purification Technology, 78 (2011) 97-102, https://doi.org/10.1016/j.seppur.2011.01.028.

8. Yonghui Wu, Jingyi Luo, Cuiming Wu, TongwenXu, Yanxun Fu, Bionic multisilicon copolymers used as novel cross-linking agents for preparing anion exchange hybrid membranes. Journal of Physical chemistry B. 115 (2011) 6474-6483, https://doi.org/10.1021/jp1122807.

9. Jingyi Luo, Cuiming Wu, Yonghui Wu, TongwenXu, Diffusion dialysis of hydrochloride acid at different temperatures using PPO-SiO2 hybrid anion exchange membranes. Journal of Membrane Science, 347 (2010) 240-249, https://doi.org/10.1016/j.memsci.2009.10.029.

10. Cuiming Wu, Yonghui Wu, Jingyi Luo, TongwenXu, Yanxun Fu, Anion exchange hybrid membranes from PVA and multi-alkoxy silicon copolymer tailored for diffusion dialysis process. Journal of Membrane Science. 356 (2010) 96-104, https://doi.org/10.1016/j.memsci.2010.03.035.

Selected Awards

• Travel support for a career trek to San Francisco, UVA, Jan. 2019

• First Place of Toastmasters International Speech Contest in Area 22, U.S., 2017

• Second Prize of the 7th Anhui Province Natural Science Outstanding Paper, USTC, 2013

• Second Prize of the 2nd Academic Annual Symposium for Graduate, USTC, 2013

• P&G Excellent Graduate Scholarship-Chinese Academy of Science, USTC, 2012

• First-authored paper “J. Membrane Sci. 347 (2010) 240-249” was honored “One of the Most Influential One Hundred International Papers in China” and recognized as “The Most Cited Authors” by Journal of Membrane Science, USTC, 2011

• Second Prize of Challenge Cup, USTC, 2011

• Outstanding Graduates-University level, HUT, 2008

• Excellent Student-University level, HUT, 2007

• Outstanding Student Scholarship-second place, HUT, 2007

• Innovation Experiment Competition-third place and University level, HUT, 2006