Catherine Yeh

CS Ph.D. Candidate @ Harvard University

Hi, I'm Catherine! I'm a third-year computer science Ph.D. student @ Harvard University, advised by Prof. Martin Wattenberg and Prof. Fernanda Viégas in the Insight + Interaction Lab. My research interests lie at the intersection of visualization, interpretability, and human-AI interaction. I'm fortunate to be supported by the NSF Graduate Research Fellowship and Kempner Institute Graduate Fellowship.

Previously, I received my B.A. in computer science + cognitive science @ Williams College. As an undergrad, I worked with Prof. Iris Howley to design an empirical framework for post-hoc explainable AI. I've also interned at Apple Research, Microsoft Research [x2], and Adobe Research. You can find more information in my CV.

Research Highlights ✨

My research explores how artificial intelligence (AI) and visualization can augment each other. Currently, I'm interested in (1) designing interactive visualization tools and methods to support model interpretability [VIS for AI] and (2) leveraging the capabilities of generative AI models to enhance visualization workflows [AI for VIS].

Filter by year:2022

2025

or tag:

Exploring Empty Spaces: Human-in-the-Loop Data Augmentation

To appear @ CHI 2025

visaihci

Exploring Empty Spaces: Human-in-the-Loop Data Augmentation

Catherine Yeh, Donghao Ren, Yannick Assogba, Dominik Moritz, Fred Hohman

We introduce Amplio, a human-in-the-loop data augmentation tool for increasing data diversity in unstructured text datasets.

paper

code

Designing a Dashboard for Transparency and Control of Conversational AI

Preprint

hciai

Designing a Dashboard for Transparency and Control of Conversational AI

Yida Chen, Aoyu Wu, Trevor DePodesta, Catherine Yeh, Kenneth Li, Nicholas Castillo Marin, Oam Patel, Jan Riecke, Shivam Raval, Olivia Seow, Martin Wattenberg, Fernanda Viégas

We show evidence that conversational language models develop internal models of their users and design a dashboard to provide transparency and control over these "user models."