Hello!
My name is Nicole Ma, and I am a freshman at Stanford University pursuing a B.S. in Computer Science and a minor in Mathematics. I'm passionate about artificial intelligence and sustainability. I'm currently developing machine learning models to enhance energy output forecasting for offshore wind applications.
Portfolio
21st Century Global and Regional Surface Temperature Projections
Used climate models to estimate calendar years of when surface temperatures will increase by 1.5, 2.0, and 2.5°C relative to the preindustrial period, both globally and in three target regions: the Arctic Circle, Tropics, and Antarctic Circle. Won the CESASC JHJ Prize for Fundamental Science in Physics.
Citation: Ma, N., Jiang, J. H., Hou, K., Lin, Y.,Vu, T., Rosen, P. E., et al. (2022). 21st century global and regional surface temperature projections. Earth and Space Science, 9, e2022EA002662.
A Novel System for Regional Twitter Hate Speech Analysis and Detection Using Deep Learning Models and Web Scraping
Constructed binary classification and multi-class CNNs (Convolutional Neural Networks) to detect hate speech from real-time Twitter data and classify tweets with hate speech into five categories. The binary classification model reached an AUC score of 98.95% and the multi-class classification model reached an AUC score of 89.46%. Won 1st in the 2022 igniteCS Programming Expo.
Citation: Ma, Nicole & Sun, Yu. (2023). A Novel System for Regional Twitter Hate Speech Analysis and Detection using Deep Learning Models and Web Scraping. 93-103. 10.5121/csit.2023.130207.
Constructing a Novel Multidirectional Stochastic Process Model to Simulate the Latitude of the North Atlantic Jet Stream
Created set of multidirectional stochastic process models to simulate the latitude of the North Atlantic Jet Stream by applying vector autoregression. Reached 36.1% improvement from the RMSE score of the baseline, unidirectional model. Won 1st Place in the Virtual Region Junior Science and Humanities Symposium (JSHS).
Using Machine Learning to Improve Offshore Wind Resource Assessment and Forecasting.
Constructed a Random Forest Regressor, a Support Vector Machine, a Convolutional Neural Network, and a Feedforward Neural Network to reduce biases in ERA5 wind speeds by over 50%. Presenting research as 1st author at the 2024 NAWEA/WindTech conference.
Experience
Machine Learning and Atmospheric Science Research Fellow
Stony Brook, Simons Summer Research Program
June 2023 - Present
- Created 8 machine learning models to bias-correct wind speed assessments and forecasts for offshore wind energy applications using TensorFlow and Python
- Reduced biases in state-of-the-art operational models by over 50% on 10 vertical levels
- Chosen as 1 of 44 research fellows from pool of ∼1000 nominees (∼4.4%)
- Abstract accepted for the 2024 NAWEA WindTech Conference; Currently first-authoring research paper
Climate Modeling Research Intern
Caltech/NASA Jet Propulsion Laboratory
Aug 2021 - Aug 2023
- Statistically analyzed 11 climate models and applied them to forecast the timing of 9 global and regional climate tipping points using Python and R
- Published research paper as 1st author in the Q1 journal Earth and Space Science; only high school 1st author out of all issues in 2022
- Presented research at the American Meteorological Society 103rd Annual Meeting
Machine Learning Research Intern
SoftCom Lab, California State Polytechnic University-Pomona
November 2021 - January 2023
- Used webscraping to collect 60K+ data points on X/Twitter
- Created two Convolutional Neural Networks to detect and classify hate speech with state-of-the-art AUC scores
- Created website using Google Firebase and Plotly to monitor >10K tweets each day and to display real-time analysis
- Published research as 1st author in the proceedings of the International Conference on Advances in Computer Science and Information Technology