Yeh Chan Yoo

This webpage goes over the research work I have done over my time as a student in University of Washington and UC Berkeley. (If you want to know more about my career, check out my LinkedIn profile or my resume!)

Graduate School Research

scVIP: personalized modeling of single-cell transcriptomes for developmental and disease phenotypes

Posted on April 22, 2026 (Link & Codebase Link)

Figure 1 from the Preprint Paper

This project was done under the leadership of Allen Institute (for Brain Science) scientist Jane Lai and the guidance of Professor Mariano Gabitto (Allen Institute profile, UW profile). The project revolves around scVIP, a generative framework that integrates transcriptional profiles and phenotypic markers to learn personalized individual-level embeddings using generative models and cell-type–aware multi-instance learning. I have contributed significantly to the documentation and codebase work for this paper, and I am named as the second author.

Note that the paper is publicly released as the preprint as of April 28, 2026.

Undergraduate School Research

CR4CR Autograder Model Presentation

Presented on October 9, 2023 (Link, PDF Link)

Illustrative image generated with ChatGPT; used solely for visual clarification and not included in the final presentation.

In this presentation, I share the results of my research as a research assistant with UC Berkeley’s Berkeley Evaluation & Assessment Research (BEAR) Center on the CR4CR project. The goal of the project was to explore how RoBERTa – a state-of-the-art large language model – could be applied to automatically grade short answers - a task with significant implications for scaling educational assessment. Through data collection, model training, and rigorous evaluation of a test set, I was able to develop a grading system that achieved a test accuracy of 75% when assessing short answers. In this presentation, I discuss the methodology, results, and limitations of the research, to further our understanding of both the potential and challenges of leveraging powerful deep learning models like RoBERTa for educational applications.

CR4CR GeoGebra Interactive Proof Presentation

Presented on October 6, 2023 (Link, PDF Link)

Animated images showing the GeoGebra demos in action

In this presentation for UC Berkeley’s BEAR Center, I discuss my past work working with GeoGebra and the potential for the program to be used in developing interactive educational material. GeoGebra is a free educational software that allows for both the creation and sharing of dynamic visualization of geometry and algebra. I discuss what the development process is like for developing GeoGebra visualizations and share my tips for developing these visualizations effectively. I also share some of the applets I made on GeoGebra, which garnered over 15,000 views overall.