Yu Duan | Tsinghua University

About Me

My name is Yu Duan (You-Dwan, in Chinese: 段昱). I joined MIT EECS as a PhD student in 2023. I am currently working with Prof. Kanaka Rajan on applying machine learning tools to model the brain. As an undergrad, I studied computer science at Institute for Interdisciplinary Information Sciences at Tsinghua University, where I worked with Prof. Kaisheng Ma and Prof. Yi Zhong on networks with synaptic plasticity and models of the fruit fly mushroom body. I was also a visiting student at Brain and Cognitive Sciences at MIT, where I worked with Prof. Robert Yang on multi-system models of working memory.

For research, I am broadly interested in using machine learning tools to advance brain sciences. Several directions that I feel excited about include

Building data-driven models that unify neural dynamics, perception, and behavior;
Developing or applying novel machine learning tools to help collect and analyze experimental data;
Explaining experimental findings in a goal-driven manner, where the goal could be modeled as loss function/reward in the supervised learning/RL framework.

In my free time, I enjoy playing piano and video games. I am a fan of both rock and classical music so I often go to concerts. Apart from that, I also love running along the Charles River.

Publications

Hebbian and Gradient-based Plasticity Enables Robust Memory and Rapid Learning in RNNs

Yu Duan, Zhongfan Jia, Qian Li, Yi Zhong, Kaisheng Ma

International Conference on Learning Representations (ICLR 2023)

PDF Code

Human-like Capacity Limits in Working Memory Models Result from Naturalistic Sensory Constraints

Yudi Xie*, Yu Duan*, Aohua Cheng, Pengcen Jiang, Christopher Cueva, Guangyu Robert Yang

Computational and Systems Neuroscience (COSYNE 2023)