It was another year of growth for the graduate certificate program at the Center for Statistics and Machine Learning (CSML) which saw 16 graduate students completing the program, an increase from 13 in the previous year and eight for the first cohort in 2019.
Of this year’s graduates: four were doctoral students from the departments of mechanical and aerospace engineering, neuroscience, civil and environmental engineering, and chemical and biological engineering; one was a master’s student in Princeton’s School of Public and International Affairs (SPIA); and 11 students earned master’s degrees in finance at the Bendheim Center for Finance.
“We are pleased with the expansion of the graduate certificate program, which shows that data science and machine learning are becoming vital tools in research and industry,” said Peter Ramadge, CSML director. “I am especially proud that these students conducted their work and completed the program while living under the constraints of the pandemic. It’s been a tough year, but they have risen to the challenge.”
For the students, the certificate program was an enriching, valuable experience that will inform their future work.
“For me, statistics and machine learning are ways to make more holistic, better informed policies,” said Anthony Cilluffo, who earned a master’s degree in public affairs at SPIA and is looking to work in public policymaking. “This is especially the case for government agencies with an expanding mission but a shrinking budget.”
His research project was using data on complaints about police misconduct to predict which police officers were most likely to commit misconduct in the future.
“I used machine learning methods, including a random forest classifier, and data on an officer's prior conduct record. If implemented, my prediction model could allow police departments to move from responding to misconduct after it happened to proactively intervening before misconduct takes place,” he said.
Francisco J. Carrilllo, earned his doctoral degree in chemical and biological engineering, and used data science and machine learning to predict stochastic clogging mechanisms such as traffic jams, pipe obstructions and trash accumulation.
“Clogging is a concept that is very easy to conceptualize but very hard to predict and understand. Machine learning was the perfect tool to try and characterize these complex, seemingly random processes,” said Carrilllo, who is slated become a postdoctoral research associate at the Department of Energy Resources Engineering at Stanford University. “The CSML program gave me a solid and fundamental understanding how our current technology-driven word works and where it might go next and, of course, helped me in my degree research.”
Yifan Wang, a finance graduate, said she will be joining Goldman Sachs as a quantitative researcher, a role no doubt where she will be using the skills she learned during the CSML program.
“It gave me the chance to learn about machine learning theories, implement machine learning algorithms and apply machine learning methods to the topics I was interested in. Nowadays, machine learning is so widely used in the financial world, and it has greatly enhanced my hard skills,” she said.
In Wang’s research, she used cross-sectional returns to make predictions on one-minute-ahead stock returns. She used several machine learning methods to make predictions and compare the prediction performance with benchmark models.
Alina Skripets, another finance student, is going into a similar line of work as a quantitative associate at Bank of America’s Quantitative Strategies Group.
For her master’s degree, she did research on the commoditization of the California water market.
“I applied machine learning algorithms to the weekly NQH2O index data to predict future values, analyzed predictability and investigate the most impactful factors that influence the changes in spot price of California water,” she said.
She too gave the CSML certificate program a ringing endorsement, saying it gave her a broad overview of machine learning applications and broadened her quantitative tool kit.
“For me, CSML certificate was a way to take a structured approach to gaining machine learning skills. Naturally, one could simply take some courses in machine learning to dive into the most alluring subjects, but the CSML program, through its theoretical and applied machine learning requirements strikes a good balance,” she said. “In addition, the independent research gave me freedom to apply knowledge which made me appreciate the full spectrum of new possibilities that machine learning tools have to offer.”