Yutong Bai
UC Berkeley
yutongbai@berkeley.edu
Bio
Yutong Bai is a Postdoc Researcher at UC Berkeley, advised by Alexei A. Efros, Jitendra Malik and Trevor Darrell. Prior to that, she obtained CS PhD degree at Johns Hopkins University advised by Prof. Alan Yuille. Her research aims to build up the AI system with less supervision and strong robustness. Explorations include representation learning, self-supervised learning, and generative modeling.
Areas of Research
- Computer Graphics and Vision
Listening to the Data: Visual Learning from the Bottom Up
Deep learning has witnessed advancements in architectures, particularly in large language models (LLMs) that exhibit a strong affinity for data. Self-supervised generative pretraining using transformers has enabled LLMs to scale up effectively in terms of model and data size, and therefore acquire diverse skills and pattern recognition abilities. Building on these advancements, the key goals of my future research are achiev- ing higher levels of general intelligence directly from raw sensory experience, akin to how humans learn. This potentially involves continuous adaptation for better generalization in dynamic real-world settings over time with multi-modal input.