Headshot of Kyle Hsu

Kyle Hsu

徐宏愷

PhD Candidate, Stanford University

kylehsu@cs.stanford.edu

I'm a final year computer science PhD student advised by Chelsea Finn and Jiajun Wu. Much of my PhD has been generously supported by a Sequoia Capital Stanford Graduate Fellowship in Science & Engineering and an NSERC Postgraduate Scholarship – Doctoral. I'm a bilingual Taiwanese-Canadian.

My recent research comprises two themes: (i) disentangled representation learning of high-dimensional data for interpretability and safety and (ii) tooling and inductive biases for robot learning in the age of foundation models. I'm actively seeking full-time AI research and engineering roles! Please reach out if you think I may be a good fit.

Previously, I studied engineering science at the University of Toronto. In my free time, I like to ski & snowboard and play Soulslike & board games.

selected publications

disentangled representation learning

Disentanglement via latent quantization
Kyle Hsu, Will Dorrell, James CR Whittington, Jiajun Wu, Chelsea Finn
NeurIPS 2023
paper poster code
Tripod: three complementary inductive biases for disentangled representation learning
Kyle Hsu*, Jubayer Ibn Hamid*, Kaylee Burns, Chelsea Finn, Jiajun Wu
ICML 2024
paper poster code
Range, not independence, drives modularity in biologically inspired representations
Will Dorrell*, Kyle Hsu*, Luke Hollingsworth, Jin Hwa Lee, Jiajun Wu, Chelsea Finn, Peter E Latham, Timothy Edward John Behrens, James CR Whittington
ICLR 2025
paper code

robot learning

Vision-based manipulators need to also see from their hands
Kyle Hsu*, Moo Jin Kim*, Rafael Rafailov, Jiajun Wu, Chelsea Finn
ICLR 2022 oral
paper project page
Evaluating real-world robot manipulation policies in simulation
Xuanlin Li*, Kyle Hsu*, Jiayuan Gu*, Karl Pertsch, Oier Mees, Homer Rich Walke, Chuyuan Fu, Ishikaa Lunawat, Isabel Sieh, Sean Kirmani, Sergey Levine, Jiajun Wu, Chelsea Finn, Hao Su, Quan Vuong, Ted Xiao
CoRL 2024
DGR@RSS2024 spotlight
paper project page

few-shot learning

Unsupervised learning via meta-learning
Kyle Hsu, Sergey Levine, Chelsea Finn
ICLR 2019
paper project page
Unsupervised curricula for visual meta-reinforcement learning
Allan Jabri, Kyle Hsu, Ben Eysenbach, Abhishek Gupta, Sergey Levine, Chelsea Finn
NeurIPS 2019 spotlight
paper project page
FSPO: few-shot preference optimization of synthetic data in LLMs elicits effective personalization to real users
Anikait Singh*, Sheryl Hsu*, Kyle Hsu, Eric Mitchell, Stefano Ermon, Tatsunori Hashimoto, Archit Sharma, Chelsea Finn
preprint
paper project page

scalable abstraction-based controller synthesis

Multi-layered abstraction-based controller synthesis for continuous-time systems
Kyle Hsu, Rupak Majumdar, Kaushik Mallik, Anne-Kathrin Schmuck
HSCC 2018
paper code
Lazy abstraction-based controller synthesis
Kyle Hsu, Rupak Majumdar, Kaushik Mallik, Anne-Kathrin Schmuck
ATVA 2019 invited paper
paper code