How can AI find patterns in complex data?
How a KEYS intern from Nogales joined University of Arizona researchers using artificial intelligence to uncover patterns in financial data.
Pablo Fernandez (right), a 2026 KEYS intern from Salpointe Catholic High School in Tucson, and faculty mentor Haiquan Li (left), associate professor of biosystems engineering, worked together on artificial intelligence research at the University of Arizona.
Caroline Bartelme, BIO5 Institute
From financial markets to medical records, enormous amounts of data are generated every day. Haiquan Li and his research team use artificial intelligence to identify patterns and solve complex real-world problems.
This summer, 2026 KEYS intern Pablo Fernandez from Salpointe Catholic High School joined the research team of Li, associate professor of biosystems engineering in the College of Engineering and College of Agriculture, Life & Environmental Sciences, to apply those AI techniques to one particularly complex challenge: predicting stock market trends.
Working alongside Li, a member of the BIO5 Institute, Fernandez analyzed computer code, trained transformer-based AI models and combined financial news with market indicators to improve the model's predictive accuracy.
For Fernandez, the project offered something he hadn't expected: the freedom to make decisions about the direction of his own research.
"What really excites me is the amount of freedom we have," said Fernandez. "I can decide where to guide the project and how to progress."
One of the biggest lessons came after Fernandez believed he had dramatically improved the model's performance. After reviewing the results with Li, he discovered the improvement came from changes in the dataset rather than the model itself. Although the breakthrough wasn't real, the experience gave him a much deeper understanding of how AI models are evaluated and improved.
Those moments are exactly what Li enjoys most about mentoring.
"You never know how the young generation thinks about a problem," said Li "Sometimes they make big breakthroughs."
By the end of the summer, Fernandez contributed to research improving an AI model designed to recognize patterns in financial markets. More than anything, the experience confirmed that he enjoys tackling open-ended problems and wants to pursue computer science as a career.