Kaya Stechly

I am a Linguistics M.A. student at Arizona State University interested in theoretical models of language learning, computational cognitive science, and language-mediated reasoning in both brains and machines. I am currently applying to PhD programs, targeting a Fall 2025 start.

My current reasearch is split into two major threads:

On the AI side, I currently work at the Yochan lab under Subbarao Kambhampati. My research is guided and dual goals of teaching concepts to and distilling symbols from AI systems in human-interpretable form. Much of my work has focused on the investigating much-hyped in-context learning abilities of large language models, especially as applied to classical reasoning and planning tasks. I also work on leveraging the strengths of neural networks in representing tacit knowledge to improve the expressivity and efficiency of symbolic systems.

In linguistics, I am studying models of phonological acquisition and evolution in the context of optimality theory and its extensions. I am particularly interested in the question of how linguistic typology is shaped not just by binary (possible vs. impossible) representational factors but by biases induced by the resource-efficiency of the learning algorithms that the human brain implements. My thesis advisor is Kathryn Pruitt.

In my free time, I enjoy swimming, reading and writing fiction, and baking.