Heng Yang

Assistant Professor@Harvard University·美国·OpenReview

研究方向

Reinforcement Learning · Conformal prediction · Computer Vision · Convex Optimization · Convex Relaxations · Semidefinite Programming and Relaxations · Robotics

4.5

合作者 (20)

Heng Yang

MoTVLA: A Vision-Language-Action Model with Unified Fast-Slow Reasoning

Accelerating Structured Chain-of-Thought in Autonomous Vehicles

LoRA3D: Low-Rank Self-Calibration of 3D Geometric Foundation models

Control-oriented Clustering of Visual Latent Representation

Learning Multiple Initial Solutions to Optimization Problems

The Benefit of Being Bayesian in Online Conformal Prediction

Cocoon: Robust Multi-Modal Perception with Uncertainty-Aware Sensor Fusion

Learning Multiple Initial Solutions to Optimization Problems

Adapting Prediction Sets to Distribution Shifts Without Labels