Tengyu Song

My name is Tengyu Song. I'm a PhD student at George Mason University. My research interests range from interpretable machine learning, robust statistics and differential privacy. Please check out my portfolio below, or visit my Blogs or Github. If you want to shoot a message, please email me at songtengyu33@gmail.com.

Résumé

Selected Projects

Dynamic Pricing for NYC Subway

Utilizing the MTA turnstile data, we built a dynamic pricing model based on ridership prediction and demand elasticity.

Learn more

EGSC Quality of Life Survey 2023

Worked with QoL Survey committee to develop the highest quality presentation of survey data for the annual Town Hall.

Learn more

Has the Left Gone Too Far? LLM-driven Insights into Think Tank Rhetoric

Utilizing GPT-4, I analyzed the change of rhetoric of Urban Institute, a centrist think tank, over the past decade.

Learn more

More Projects / Presentations to Explore

Mutual information, Fairness and Robustness in AI

Further study based on the paper "FR-Train" (Y. Roh et al., 2020): A mutual information based approach to train fair and robust models.

Learn more

R Bootcamp for TLL AT Columbia Psychology

Developed a 6-course R Bootcamp for the TLL at Columbia Psychology. The course covers basic R programming, data wrangling, data visualization, and statistical modeling.

Learn more

PV Cell Panel Defect Detection

Devised a novel algorithm to detect defects on PV cell panels using image processing and deep learning techniques.

Learn more