News
Research updates, paper releases, and recent milestones.
Recent Updates
A lightweight timeline of recent papers, research directions, and project milestones.
-
Feb. 2026
Released ALOE on action-level off-policy evaluation for vision-language-action model post-training. -
Nov. 2025
Released Discover, Learn, and Reinforce, a framework for scaling VLA pretraining with diverse RL-generated trajectories. -
Sep. 2025
Released Beyond Human Demonstrations, exploring diffusion-based reinforcement learning for VLA training data generation. -
2025
Expanded my research from general reinforcement learning toward embodied AI, VLA training, and real-world robotic system capability. -
2025
Contributed to projects on VLA inheritance, latent action modeling, and robust embodied manipulation. -
2024
Co-authored ICML 2024 work on constrained ensemble exploration for unsupervised skill discovery. -
2023
Published Behavior Contrastive Learning for Unsupervised Skill Discovery at ICML 2023.
This page can be updated over time with new papers, project releases, invited talks, awards, internships, and other milestones.