Junxuan (Helen) Shen

Hi! I’m Junxuan, a PhD student in EECS at MIT advised by Professor Yilun Du. I work at the intersection of generative modeling, robot learning, and planning. I am particularly interested in developing compositional methods that improve robots’ abilities to reason about the physical world and solve long-horizon tasks. My broader goal is to make learning-based robot systems more reliable for industrial and field applications.

Robot learning research

Hybrid Reasoning through Compositional Structural Energy Minimization
Junxuan Shen, Haonan Chen, Weirui Ye, Yilun Du
Under review.

Robot Learning with Visual Predicted Force
Haonan Chen*, Feiyang Wu*, Yuxiang Ma, Mustafa Mete, Pengfei Ye, Junxuan Shen, Cheng Zhu, Aurora Ruggeri, Kelvin Cheung, Jiayuan Mao, Edward Adelson, Jiajun Wu, Robert D. Howe, Yilun Du
* Equal contribution.
Preprint, 2026. Paper · Project page

See my papers and preprints for the full list.

Before robot learning

I received my MS in EECS from MIT in May 2026, advised by Professor Thodoris Lykouris. My master’s research focused on theoretical aspects of online decision-making and online learning, including learning-augmented scheduling and Bayesian optimal stopping.

Before MIT, I earned my BS in computer science and mathematics from Caltech in 2024. I was very fortunate to work with Professors Adam Wierman, Guannan Qu, Nicolas Christianson, and Adam Sheffer on learning-augmented algorithms, control, and extremal combinatorics.