Mary Gates Research Scholar, Autumn 2024
Research Project: Order-Preserving Dimension Reduction for Multimodal Semantic Embedding
Project Description: I developed a method called Order-Preserving Dimension Reduction (OPDR) to make multimodal data retrieval more efficient. The goal was to reduce the size of high-dimensional embeddings while preserving the ranking of similar items for K-nearest neighbor (KNN) search. Our method significantly cuts down computation time without sacrificing retrieval accuracy and can be applied to a wide range of datasets and models.
What have you learned throughout the research project?
Doing research often comes with unexpected challenges—experiments may fail, and papers may get rejected. However, I’ve learned that persistence and adaptability are key. Every setback is a learning opportunity, and staying curious and open to feedback helps me grow as a researcher.
What piece of advice do you have for future applicants?
Don’t give up, even when things get tough. Research and leadership projects often come with setbacks, but persistence, curiosity, and a willingness to learn from failure will carry you far. Believe in your growth and keep moving forward!