Helen Martinez | Undergraduate Research Assistant

Helen Martinez

Undergraduate Research Assistant

Based in USA

About

Undergraduate Student @ University of Nebraska-Lincoln majoring in Mathematics with concentrations in Statistics & Data Science

Skills

Projects

Change in Economic Output of Asia Visualization

Visualizing the statistics of how Asia's economy changed over a period of time

Dataset Exploration of World Development Indicators Using SQL

Exploring the GDP of countries using SQL Queries

Visualizing the Statistics of the Best-Selling Books

Used Python to visualize a dataset that included the best-selling books of all time.

Early Forecasting of Student Performance in STEM Using Language Models Poster

Worked with Dr. Mohammad R. Hasan & Ahatsham Hayat @ University of Nebraska-Lincoln College of Engineering. We assess how well LMs process and learn from a natural language dataset that includes student grades & student background information to predict academic performance early in the semester.

Early Forecasting of Student Performance in STEM Using Language Models Paper

The final research paper for the McNair Summer Research Experience assesses how well LMs process and learn from a natural language dataset that includes student grades & student background information to predict academic performance early on in the semester.

Experience

Dec 2023 - May 2024

Undergraduate Research Assistant

- Cleaned, processed, and structured a 48-dimensional dataset to ensure quality input for machine learning models. - Evaluated the performance of three conversational LLMs (ChatGPT-3.5, ChatGPT-4.0, Gemini) across datasets with varying lengths (2-week, 4-week, 8-week) and feature combinations. - Performed ensemble learning experiments for early forecasting of student success.

Jun 2024 - Jul 2024

McNair Summer Researcher

- Preprocessed and transformed qualitative and quantitative data using Python and Scikit-Learn for machine learning pipelines - Improved model accuracy by integrating experiential and non-experiential data, outperforming non-language models. - Conducted 6 experiments analyzing cognitive, non-cognitive, and distal data through time-series methods. -Collaborated with faculty and peers, presenting research at the McNair Summer Colloquium (UNL & UCLA) -“Early Forecasting of Student Performance in STEM Using Language Models”

Aug 2024 - Present

Undergraduate Researcher

- Conduct 9+ experiments with language models on Python to evaluate model performance -Models: Mixtral, Flan T5 - Collecting data from participants on student motivation, engagement, and academic performance. - Support participants in troubleshooting issues with a custom data-collection mobile app.