Selected research on uncertainty-aware vision-language-action models, reinforcement learning, human feedback, safety, multimodal alignment, and human-centered autonomy.
Latest profile: Google Scholar. * indicates equal contribution.
arXiv preprint arXiv:2606.18043, 2026
Introduces Velocity-Field Disagreement (VFD) for epistemic uncertainty in flow-based VLAs and SAVE, an uncertainty-guided active fine-tuning framework that needs at least 22% fewer expert demonstrations than baselines.
In IEEE International Conference on Robotics and Automation (ICRA), 2026
Reframes preference-based RL as multi-objective learning: language explanations are parsed into objective-specific labels, weights, and highlights for scalarized policy optimization.
In International Conference on Machine Learning (ICML), 2026; also accepted to the ICLR 2026 Workshop on Scaling Post-training for LLMs (SPOT)
Introduces Self-Distillation Policy Optimization (SDPO), converting rich environment feedback into dense self-distillation signals for more sample-efficient RL with language models.
In IEEE International Conference on Robotics and Automation, 2024
Improves preference-learning sample efficiency by augmenting human feedback with synthesized preference queries from latent interpolation.
In IEEE International Conference on Robotics and Automation, 2024
Combines preference feedback with temporal highlights to improve granularity and representation learning. Nominated for Best HRI Paper, Best Student Paper, and Best Conference Paper.
In ACM/IEEE International Conference on Human-Robot Interaction, 2024
Uses zero-shot language-model reasoning over optional textual descriptions to align learned rewards with human preferences.
In WACV LLVM-AD Workshop, 2024
Explores few-shot LLM reasoning for inferring autonomous-system requirements from in-cabin natural-language commands. Best Student Paper Award.
In IEEE/RSJ International Conference on Intelligent Robots and Systems, 2023
Proposes a VAE-based active-learning strategy for diverse and informative preference-query selection.
In IEEE International Conference on Robotics and Automation, 2023
Narrows policy search with baseline objectives and requests human feedback when preferences matter most.
In IEEE Robotics and Automation Letters, 2021
Learns shield parameters from human feedback to obtain policies perceived as safe.