Michael H. Lim
(Hyun Jae Lim)


UC Berkeley - PhD in Statistics & Robotics
Harvard University Class of 2018
"ML does ML"

I am a 3rd year PhD student in Statistics with a focus on robotics at UC Berkeley, grateful to be advised by Prof. Claire J. Tomlin and Prof. Zachary N. Sunberg. I am interested in the intersection of control theory, sequential decision making, and machine learning, with applications in robotics.

My research goal is to develop algorithms and systems that enable physical robots to operate intelligently, safely, and efficiently, with provable guarantees. Currently, my research trajectory consists of two general approaches:

1. My research aims to develop MDP and POMDP algorithms that tackle realistic assumptions and frameworks - such as having continuous observation and action spaces - which combine provable theoretical convergence guarantees as well as computational efficiency.

2. My research attempts to interface controls and planning techniques with computer vision and machine learning, which enables agents to navigate in unknown environments or in the presence of other agents with unknown intentions.


April 2020

  • Our work on POMDP convergence guarantees for continuous observations has been accepted to be published at IJCAI-PRICAI 2020. Preprint version is available on arXiv link.
  • I will be interning at nuro.ai over the summer, working with the control and planning team under SWE/research role.