Boyi Li

Research Abstract:

The computer vision community has embraced specialized models trained on fixed object categories like ImageNet or COCO. However, relying solely on visual knowledge may limit flexibility and generality, requiring additional labeled data and hindering user interaction. My research aims to develop intelligent systems that are efficient and learn not only from visual perceptions of the physical world but also from their interactions with humans, incorporating elements such as language. These systems are intended to perform a variety of complex tasks, aiding humans and facilitating fluid interactions between humans and computers in both virtual settings and real-life scenarios. Specifically, my work has made key contributions to addressing several fundamental challenges, including open-vocabulary visual recognition, reliable content creation and interactive task planning. By addressing these challenges, I am working toward realizing the vision of creating intelligent systems capable of perceiving, reasoning, and acting in ways analogous to human beings.

Bio:

Boyi Li is a Research Scientist at NVIDIA Research and a Postdoctoral Scholar at Berkeley AI Research, advised by Prof. Jitendra Malik and Prof. Trevor Darrell. She received her Ph.D. at Cornell University, advised by Prof. Serge Belongie and Prof. Kilian Q. Weinberger. Her research interest is in computer vision and machine learning. In particular, her vision is to develop intelligent systems that are efficient and generalizable, capable of learning not only from visual perceptions of the physical world but also from their interactions with humans, incorporating elements such as language. Beyond research, she has also served as co-general chair for the Women in Machine Learning Workshop 2021 and the Women in Computer Vision Workshop 2020, 2021.