Rice graduate Andrew Lan, second-year doctoral students in electrical and computer engineering (ECE), Lucy Liu and Jack Wang, and faculty member Richard Baraniuk make up one of four teams to receive a $15,000 grand prize for their submission to the first automated scoring challenge presented by the National Center for Education Statistics (NCES).
NCES is a branch of the U.S. Department of Education that collects and analyzes data related to education in the U.S. and other nations. With a Congressional mandate, it develops comprehensive statistical data and reporting services.
The challenge was launched by the NCES in an effort to introduce data science and machine learning into operational activities and determine to what extent automated scoring programs can produce reliable scores at a reasonable cost for National Assessment of Educational Progress (NAEP) reading items.
The challenge included two components. Competitors had the option to address one or both. The first component was item-specific. In order to be successful, participants had to build a predictive model for each item that can be scored using state-of-the-art practices in automated scoring.
The second component was generic. Applicants had to build a generic scoring model that can score items not in the training dataset, but at the same grade level and subject. Lan and his team opted to focus on the first component.
Lan’s team used advanced language processing methods that reduced scoring costs without sacrificing accuracy. Lan, assistant professor of computer science at the University of Massachusetts at Amherst, earned his Ph.D. from Rice in ECE in 2016.
He was advised by Richard Baraniuk, the C. Sidney Burrus Professor of ECE. Lan, Liu and Wang met in the Digital Signal Processing group at Rice.