Computer Sciences faculty member Sharon Li has recently won National Science Foundation (NSF) CAREER Awards. Li joined the CS faculty in 2020 so is considered early in her career (pre-tenure). The NSF intends the awards to help early-career faculty build a firm foundation for a lifetime of leadership in integrating education and research.
Li’s award is part of the Robust Intelligence program at NSF and entitled “Foundations of Human-Centered Machine Learning in the Wild.” Li and colleagues’ objective is to “lay new foundations for safe, adaptive, and long-term beneficial learning algorithms in open-world environments.” These three characteristics are uniquely motivated by the symbiotic relationship between machine learning and humans in the open world. Li describes the first as “strong safety” (i.e., awareness of unknowns), which can directly inform and protect end-users who are dependent on the model’s decision. Second, humans can provide strategic inputs to guide rapid model adaptation, which is crucial in the open world with emerging new contexts. Lastly, a deep understanding of long-term dynamics allows for optimizing the benefits of model operation in the long run.
The project will directly impact real-world areas, including autonomous transportation, healthcare, commerce, and scientific discovery and includes three goals: (1) create a new learning framework for reliable decisions, rendering strong safety against unknowns upon deploying machine learning models in the wild; (2) accelerate model adaptation, learning to classify new concepts emerging in the wild while minimizing human supervision required; (3) characterize and understand dynamics in terms of long-term accuracy and safety, maximizing the impact of models as they evolve and operate in the long run.
The education plan will publicize the power of open-world machine learning through a new course, a new undergraduate mentorship program, and outreach efforts. The goal is to train the next-generation workforce to be better prepared for the inevitable paradigm shift from closed-world to open-world machine learning. “As students enter the workforce, we expect that they will adapt this knowledge to new and exciting domains, thus affecting machine learning development as well as society at large,” Li says.
Li has grand plans for the future of Machine Learning: “My long-term research vision is to transform the field of machine learning from a closed-world to an open-world setting — the former has been the focus in the past 50 years and the latter will be the new norm in the decades to come.”
Congratulations to Li on this huge accomplishment!