WRITTEN BY: Matt Busekroos
Prior to starting her PhD at CSAIL, Soya Park finished her undergraduate program at KAIST, a national research university located in South Korea. She now works with Professor David Karger in the Haystack Group, which blends approaches from human-computer interaction, social computing, databases and information management.
Park studies the way people can teach machines using their own common sense and in their own natural way. Likewise, she builds systems to facilitate such teaching.
Park said our daily life is impacted by complex intelligent systems, such as AI and ML, however, not everyone understands how technology works. As a result, non-tech-savvy users are highly dependent on engineers to build intelligent systems.
While the users of these intelligent systems are technical experts in their own domains, they do not have the knowledge to train and improve the models themselves, according to Park. This creates a disconnect: these users are the key stakeholders of their system, but they must rely on third parties to make any improvements related to performance.
“In my research, I'm hoping my work can bridge gaps between domain experts and engineers in intelligent systems development workflow,” Park said. “I approach this by building a user interface where domain experts can explore training data and model results together with engineers.”
Park said her advisor Karger was ultimately instrumental in her coming to CSAIL. She calls him her role model and has learned a lot from him in her time at the lab.
“During work time, he is a passionate and inspiring researcher,” she said. “After work time, he is a joyful dancer. He has such a great work-life balance, which I think is a very important virtue of a researcher. I often tire myself too much and don't give myself a break. But since researching is a long-term marathon, I should learn my pace and balance research and fun.”
Park said she loves research and plans to continue with it after graduation next year. She said she enjoys working in academia, but also admires the access to real use cases and users that you get in industry.
Park is currently building an interface to facilitate collaboration between stakeholders and machine learning engineers. She encourages you to reach out if you are interested in trying out her tool in your workflow.
You can follow Park on Twitter (https://twitter.com/park_soya). She frequently posts updates about research and her dog, Toby!