Cherrie Chang, PhD

Cherrie is an aspiring computational cognitive scientist currently driven by this question: when we build abstractions to understand complex systems—whether minds, machines, or the world—what determines which details we attend to, and what does this reveal about how we want to understand? She is curious about how the cognitive process of conceiving abstractions interacts with social inference and subjective values. In practice, this translates to an interest in using large-scale behavioral experimentation and computational modeling to explore the dynamics between utility, elegance, and validity in how we construct understanding.

She graduated with a B.A. in Cognitive Science and Computer Science with a minor in Analytic Philosophy from Vassar College. There, She compared the representational structures underlying human and artificial cognition by probing whether large language model embedding spaces structurally capture human semantic relationships, revealing meaningful differences in how AI systems organize conceptual knowledge. This work reflects her broader interest in understanding minds through their internal representations and the abstractions we use to make sense of them.

Here at IHP’s Advanced Analytics & AI for Communication Science Lab led by Dr. Josh Hartshorne, She is using neural networks to investigate how age of language acquisition affects accent. She also works on developing Pushkin, a research software that aims to provide end-to-end infrastructure for behavioral scientists to deploy and run experiments at scale. This line of work continues from her previous position at Vassar College, where she worked on developing jsPsych, a popular JavaScript library for building behavioral experiments. Her work positions her at the intersection of building experimental tools and understanding the cognitive phenomena they measure—experience that informs both her technical work and her theoretical questions about how abstractions shape understanding and, by extension, scientific practice itself.