USC Undergraduate Research Associates Program (URAP)

Available Fall 2023-Spring 2024 semesters for 8-10 hours per week
Compensation: $1500/semester ($3000 total)
How to apply: Fill the interest form linked above.
Eligibility: Undergraduate rising sophomore/junior/senior who does not hit any ineligibility criteria below. Undergraduate students are ineligible for URAP funding if they:

• Enroll in a Progressive Degree Program (PDP)
• Enroll in a 490 (directed research class)
• Officially graduated from USC
• Take a leave of absence.
• Enroll in less than 12 units at the time when you make payment request.
• Receive either a Provost, CURVE, SOAR, or SURF funding (students cannot receive two sources of funding in the same term).

Pre-requisite skills:

• Algorithms: CSCI 270; Linear Algebra: MATH 225; Probability: EE 364 or MATH 407 or BUAD 310
• Familiarity with Python and associated ML frameworks such as NumPy, Pandas, and Pytorch
• Willingness to work across multiple subfields of AI (e.g., language, vision, robotics) and read papers from out-of-discipline
• Bonus: CSCI 467 or other machine learning course (e.g., CSCI 499's that involve ML or NLP), possibly concurrently out-of-discipline with project

Skills you will acquire:

• Training and evaluating deep learning models in Pytorch
• Wrangling large scale data from multiple sources
• Designing experiments for human subjects
• Graphics and animation and/or basic robotics controls

Brief Project Description:
Human cognition is situated in the physical bodies and spaces people occupy, yet automated instruction (e.g., recorded videos of classes or tutorials) is typically communicated verbally to stationary learners. This work will leverage embodied tutors, from virtual avatars to physical robots, to engage learners via combinations of verbal and gestural cues, and in turn study how learner gesture repetition can serve to improve information recall. Specifically, this work will test the hypothesis that learners who see a tutor perform iconic gestures representing new words that they hear from the tutor will be better able to recall those words than learners who do not see such gestures and than learners who do not repeat both the verbal and gestural inputs to the tutor.
Experiments will be conducted in the context of second language (L2) acquisition, where an ongoing subtask is lexical acquisition: memorizing content words in the target language. Prior work has shown that, in the context of second language (L2) learning, both listening comprehension and word retention are improved when instruction is accompanied by tutor gestures. This work will evaluate the hypothesis that combining iconic and metaphoric gesture production during both speaking and hearing of information those gestures represent will improve retention of new information for learners, and that this effect will be stronger with a robot tutor than with recordings or avatars.

CURVE Fellowship

Available Fall 2023-Spring 2024 semesters for 8-10 hours per week
Compensation: $1250/semester ($2500 total)
How to apply: Program details, including how to apply, are available at the CURVE program page.
Eligibility: All CURVE applicants must meet the following eligibility requirements:

• Be enrolled as full-time students (12 units minimum) for the semester in which they receive an award.
• Be declared a major in one of the degree programs in the Viterbi School of Engineering.
• Students of any citizenship status are eligible to participate (however, please note that participation in certain research projects may require U.S. citizenship)
• Demonstrated ability to commit required weekly hours to research
• For continuing researchers, a minimum cumulative GPA of 3.0 or above.

Project Information: On the CURVE website, available projects are grouped by department and lab. Under the Thomas Lord Department of Computer Science, you can find projects that are offered by the GLAMOR Lab, alongside project descriptions, student responsibilities, and other information.