Scholar
Kyle Hsu
Google Scholar ID: KCdL5B0AAAAJ
Stanford University
Citations & Impact
Citations
9,226
H-index
13
i10-index
14
Publications
18
Co-authors
14
list available
Publications
Resume (English only)
Academic Achievements
- Disentanglement via latent quantization, NeurIPS 2023
- Tripod: three complementary inductive biases for disentangled representation learning, ICML 2024
- Range, not independence, drives modularity in biologically inspired representations, ICLR 2025
- Flow to the mode: mode-seeking diffusion autoencoders for state-of-the-art image tokenization, ICCV 2025
- Vision-based manipulators need to also see from their hands, ICLR 2022 oral
- Evaluating real-world robot manipulation policies in simulation, CoRL 2024, DGR@RSS2024 spotlight
- Unsupervised learning via meta-learning, ICLR 2019
- Unsupervised curricula for visual meta-reinforcement learning, NeurIPS 2019 spotlight
- FSPO: few-shot preference optimization of synthetic data in LLMs elicits effective personalization to real users, preprint
- Multi-layered abstraction-based controller synthesis for continuous-time systems, HSCC 2018
- Lazy abstraction-based controller synthesis, ATVA 2019 invited paper
Research Experience
- Senior ML Engineer at Tesla, working on AI for the Optimus project.
Education
- PhD from Stanford University, advised by Chelsea Finn and Jiajun Wu; supported by a Sequoia Capital Stanford Graduate Fellowship in Science & Engineering and an NSERC Postgraduate Scholarship – Doctoral.
Background
- Research interests include representation learning, robot learning, and few-shot learning. Specializes in AI, particularly working on the Optimus project at Tesla.
Miscellany
- Bilingual (Taiwanese-Canadian), enjoys skiing, snowboarding, playing Soulslike games, and board games in his free time.
