Research Experience for Undergraduates
Stanford’s School of Engineering is home to amazing researchers and research.
As an undergraduate, you can be a part of it through the Research Experience for Undergraduates (REU) program, which provides opportunities to work on cutting-edge research guided by Stanford faculty. Professors are excited to work with bright, enthusiastic undergraduates.
Information about the program will be sent in early January to declared MS&E majors and minors, and students who have begun the declaration process in Axess.
The Department of Management Science and Engineering seeks applications from current Stanford undergraduate students in early February each year and matches applicants with participating faculty members. Students participate in 10 weeks of research with a faculty member, culminating in a presentation to all participants and participating faculty. Weekly group meetings are held with all participants.
Any other interested students may contact Lori Cottle in January. For general advice on how to get involved in research at Stanford, refer to the Stanford Undergraduate Advising and Research (UAR) Guide.
See the kind of work REU students are doing below:
Summer 2021 REU projects
During the last week of summer quarter, REU participants presented their work virtually. Three of those presentations are described below:
Investigating value dissemination in tech companies
Nina Prabhu (Computer Science)
Rick V. (Mathematical and Computational Science)
Many companies have values that are integral to their company culture and their products. We investigated how companies can measure whether these values are being internalized and communicated by their employees.
Tools and methodologies used
- Glassdoor review scraping
- Sentiment analysis models
- Python data analysis
- Data visualization
- Sustainability Accounting Standards Board (SASB) materiality map—provides insight as to what metrics/values to track across companies
We found that analyzing public documents (letter to shareholders, sustainability reports), may be indicative of employee sentiment at larger tech companies. By using a sentiment analysis model and identifying value-oriented words and tracking their occurrences, we also found that employees are not primarily motivated by the core values of their company, indicating that these values might not be currently communicated enough during the hiring process and on the job daily.
By increasing their emphasis on values throughout the hiring process and on the job, founders and CEOs at small and large companies alike can create a more cohesive and value-oriented culture. For larger companies, this can create a better sense of community, while at smaller companies, this will set founders up for success as their company grows.
Predicting undermatching in college admissions
Sajel Galhotra (Mathematical and Computational Science)
Research shows that selective colleges have better educational outcomes, but many students only apply to universities of lower selectivity than they are qualified for, a phenomenon we call "academic undermatch". Additionally, many students do not apply to schools with the lowest possible cost for a given selectivity level, a phenomenon we call "financial undermatch". Data from the Common Application shows that about 40% of students academically undermatch and about 70% of students financially undermatch. We want to intervene before a student undermatches.
Tools and methodologies used
In order to perform these interventions, we built numerous prediction models of varying complexity to predict as early as possible which students are most likely to undermatch.
We found that we can predict undermatch as early as the beginning of a student's junior year of high school, but financial undermatch is much easier to predict than academic undermatch.
We can use these models to identify students at high risk for undermatch and target our interventions to these students so that they can access the better educational outcomes offered by selective colleges.
Simulation of max-stable random fields and estimation of their densities
Jason Zhu (Mathematics; coterminal Master's candidate in Computational and Mathematical Engineering)
We are interested in the computational and statistical techniques used in the modelling of rare, but consequential, events, e.g. extreme precipitation, abnormally high temperatures, or failure of machinery components. Thus, rare event modeling is of paramount importance when it comes to understanding how to allocate resources to alleviate or avoid extreme circumstances that may put people at risk. In standard statistical settings, the central limit theorem gives that sums of independent statistical errors is approximately normal for large sample sizes. In the case of modeling extremes, instead of taking sums of statistical errors, we take the maxima of our sample measurements, so we must use a different statistical model called max-stable distributions. Thus, performing statistical estimation of rare events coincides with the problem of density estimation of these max-stable distributions.
Tools and methodologies used
In this project, we developed Monte Carlo estimators for max-stable distribution densities. This first involves the simulation of these max-stable random fields, which in turn, involves the simulation of a series of random times. These random times are developed such that the relevant quantities we are interested in measuring can be simulated in an almost surely finite amount of time. Afterwards, we apply multi-level Monte Carlo methods to generate a finite-variance, unbiased estimator of the joint density of the random field.
We can use our estimators as a tool to help decide where, when, and how to allocate resources, e.g. to respond to extreme weather events like floods.
See brief descriptions of past REU presentations here.