Claire Jin wants to improve how generative artificial intelligence is incorporated into robotics.

The third-year student in the School of Computer Science was selected as one of CMU's three 2024 Goldwater Scholars, one of the most prestigious STEM scholarships for undergraduates. This award comes from the federally endowed Barry Goldwater Scholarship and Excellence in Education Foundation. It provides up to $7,500 per academic year, covering tuition, books, fees, and room and board. 

"Knowing the caliber of previous scholars and their remarkable achievements in both industry and academia, I feel a sense of awe and gratitude," Jin said. "It's an incredible honor to be recognized among such esteemed individuals, and I'm eager to contribute similarly with my future research."

Jin is working with SCS Assistant Professor Yonatan Bisk in the Connecting Language to Action and the World (CLAW) Lab on how AI programs reason and converse when they collaborate with humans or other robots to perform tasks without a single solution. She wants to help create adaptable, reliable and transparent robotic systems that integrate with generative AI. By incorporating physical intuition, situational awareness and the ability to infer human intentions, Jin wants to improve how people interact with robots. Additionally, by improving model generalizability to a variety of tasks and interpretability, she aims to help users troubleshoot more easily.

Jin, who will earn her bachelor's degree in AI with an additional major in statistics in December, already has a great deal of experience working with large language models (LLMs) and plans to pursue a Ph.D.

Her previous work includes using LLMs for text-based game development and bug detection with Microsoft Research's Natural Language Processing Group, refining multimodal models with its Knowledge and Language Team, and optimizing queueing strategies through reinforcement learning with Assistant Professor Weina Wang. She co-authored a paper that was accepted to the Conference on Computer Vision and Pattern Recognition and one that was accepted to the IEEE Conference on Games. Recently, she submitted a paper on game bug detection using LLMs to the Annual Meeting of the Association for Computational Linguistics as the first author.

Jin began her research journey in high school, where she led a project with mentorship from Zita Oravecz in the Department of Human Development and Family Studies at Penn State University on predicting stress onset in adolescents based on physiological signals collected using wearable devices. Jin later published this work as the first author in 2023, following reanalysis using deep learning methods Jin learned at CMU, and presented it at the IEEE-EMBS International Conference on Body Sensor Networks with support from the Carnegie Mellon OURSD Presentation Award.

A 2023-24 Andrew Carnegie Society Scholar, Jin has been actively involved in the wider campus community. She is an SCS representative to the Undergraduate Student Senate, serving as the business affairs committee chair for the past two years, and she formerly served as the vice chair of the joint funding committee. She has also been a residential advisor for the past two years in the Henderson and Morewood Gardens dorms.

SCS has played a pivotal role in shaping Jin's career, from the classes she's taken to the mentorship she's received from her academic and research advisors — Bisk, Wang and Reid Simmons —as well as professors Daniel Sleator, Daniel Anderson and David Woodruff, to whom she's been a teaching assistant.

"Their mentorship has not only fostered my academic growth but has also imparted valuable lessons that have shaped me into a more well-rounded individual," Jin said.

You can learn more about the scholarship on the Goldwater Foundation website.