5 papers · 1 filter
Sampling-based Kinodynamic Planning with World Models
Yitian Gao, Hanlan Yang, Zachary Kingston
Visual world models predict the consequences of actions from images, but planning with them is typically done by optimizers such as CEM and MPPI. However, these optimizers stall on…
Safe Real-Time Policy Steering via Noise-Space Trajectory Optimization for One-Step Generative Policies
Qingyi Chen, Joseph Ruan, Zachary Kingston
Generative robot policies can represent diverse, multimodal behaviors, but adapting pretrained policies to deployment-time constraints such as collision avoidance and orientation m…
Revisiting Replanning from Scratch: Real-Time Incremental Planning with Fast Almost-Surely Asymptotically Optimal Planners
Mitchell E. C. Sabbadini, Andrew H. Liu, Joseph Ruan +3
Robots operating in changing environments either predict obstacle changes and/or plan quickly enough to react to them. Predictive approaches require a strong prior about the positi…
CDE: Concept-Driven Exploration for Reinforcement Learning
Le Mao, Andrew H. Liu, Renos Zabounidis +3
Intelligent exploration remains a critical challenge in reinforcement learning (RL), especially in visual control tasks. Unlike low-dimensional state-based RL, visual RL must extra…
Parallel Heuristic Search as Inference for Actor-Critic Reinforcement Learning Models
Hanlan Yang, Itamar Mishani, Luca Pivetti +2
Actor-Critic models are a class of model-free deep reinforcement learning (RL) algorithms that have demonstrated effectiveness across various robot learning tasks. While considerab…