Sample-efficient Cross-Entropy Method for Real-time Planning

08/14/2020

∙

Trajectory optimizers for model-based reinforcement learning, such as the Cross-Entropy Method (CEM), can yield compelling results even in high-dimensional control tasks and sparse-reward environments. However, their sampling inefficiency prevents them from being used for real-time planning and control. We propose an improved version of the CEM algorithm for fast planning, with novel additions including temporally-correlated actions and memory, requiring 2.7-22x less samples and yielding a performance increase of 1.2-10x in high-dimensional control problems.

READ FULL TEXT

Sample-efficient Cross-Entropy Method for Real-time Planning

Sample-efficient Real-time Planning with Curiosity Cross-Entropy Method and Contrastive Learning

Model-Predictive Control via Cross-Entropy and Gradient-Based Optimization

CEM-GD: Cross-Entropy Method with Gradient Descent Planner for Model-Based Reinforcement Learning

Cross-entropy method in application to SIRC model

An Online Prediction Algorithm for Reinforcement Learning with Linear Function Approximation using Cross Entropy Method

Continuous Trajectory Planning Based on Learning Optimization in High Dimensional Input Space for Serial Manipulators

Planning with Pixels in (Almost) Real Time

Sample-efficient Cross-Entropy Method for Real-time Planning

Related Research

Sample-efficient Real-time Planning with Curiosity Cross-Entropy Method and Contrastive Learning

Model-Predictive Control via Cross-Entropy and Gradient-Based Optimization

CEM-GD: Cross-Entropy Method with Gradient Descent Planner for Model-Based Reinforcement Learning

Cross-entropy method in application to SIRC model

An Online Prediction Algorithm for Reinforcement Learning with Linear Function Approximation using Cross Entropy Method

Continuous Trajectory Planning Based on Learning Optimization in High Dimensional Input Space for Serial Manipulators

Planning with Pixels in (Almost) Real Time