The MultiComp Lab has a great tradition of including undergraduate and professional masters students in our research. The opportunities listed at the left are single-term, unfunded research appointments that may be extended to additional semesters. Each also includes a brief description, required skills, and contact information. Please email Nicki Siverling with any questions.
Opportunities for the 2020 Summer Term include:
Modeling subjectivity in human judgments
Mentor: Victoria Lin
Description: Subjectivity uncertainty is uncertainty that arises from data with subjective or ambiguous labels—labels on which human annotators themselves can disagree. This type of uncertainty can be highly informative; for example, a facial expression that is ambiguously sad is different from a facial expression that is unambiguously sad, and consequently, models should treat them differently. It is therefore beneficial to be able to explicitly quantify and characterize subjectivity uncertainty. In this project, we will design a study to collect subjective data; experiment with ways to disentangle subjectivity uncertainty from other types of uncertainty, such as observation error; and develop models that leverage subjectivity uncertainty to improve predictive performance and explainability.
Skills/Experience: Candidates should have a strong background in math and statistics and be comfortable coding in Python and its statistics and machine learning libraries (e.g. statsmodels, sklearn). Prior experience with deep learning libraries like PyTorch is a plus. Candidates with equivalent ability in R will also be considered.
Contact: Interested students should send an email to (Victoria Lin) with their CV.