CEM
A sampling-based optimization method that searches for optimal action sequences by iteratively updating the distribution of candidate solutions. It is commonly used for planning in reinforcement learning.
Timeline (last 90 days)
In model-based reinforcement learning, CEM and MPC planners work by sampling, evaluating, and comparing candidate action sequences on a learned dynamics model.
The authors explicitly state that it has not yet been established whether these local action geometry deviations directly cause planning failures, which is the question to be answered next.
The authors are interested in incorporating local action geometry into CEM's sampling covariance to improve exploration and control performance, but whether standard CEM's covariance adaptation mechanism already implicitly compensates for geometric deviations remains undetermined.
All Facts (3)
In model-based reinforcement learning, CEM and MPC planners work by sampling, evaluating, and comparing candidate action sequences on a learned dynamics model.
50%UnverifiedThe authors explicitly state that it has not yet been established whether these local action geometry deviations directly cause planning failures, which is the question to be answered next.
50%UnverifiedThe authors are interested in incorporating local action geometry into CEM's sampling covariance to improve exploration and control performance, but whether standard CEM's covariance adaptation mechanism already implicitly compensates for geometric deviations remains undetermined.
50%