GASP: The Graduate Admission Support Program

Osvaldo Banuelos, Jeff Mahler, Tyler McDonnell, Aaron Sanchez, and Stephen Walker


Background

Many graduate programs, especially in the STEM fields, have been receiving an increasing number of applicants in recent years. The traditional review process leaves committees, typically composed of departmental professors with many other commitments to research and current students, with prohibitively large numbers of applicants to review.

The System

GASP is an experimental statistical machine learning system developed for the Electrical & Computer Engineering Department at the University of Texas at Austin. GASP uses historical department admissions data to build a statistical model that characterizes new applicants by their likelihood of acceptance. Additionally, it provides visualizations that highlight the strengths and weaknesses of each candidate relative to the applicant pool. GASP allows committees to focus on applicants near the decision boundary and make informed and holistic decisions.

Related Work:

GRADE: Machine Learning Support For Graduate Admissions