JuliaPOMDP / JuliaPOMDP/DeepQLearning.jl

Action masking (feature request)

Open
#77 7 comments 0 reactions 0 assignees View on GitHub
Dominant language
Julia
Stars
77
Forks
14
PR merge metrics
No merged PRs in 30d

Description

For example, consider a 2D maze environment. The state is the grid coordinates, and the actions are moving left, right, up or down, but the only valid actions are those that do not cross the wall in the maze. It would be nice if the user could specify which actions are valid. The invalid actions are "masked" and ignored by the `NNPolicy` action selection.

Contributor guide

No contributing guide indexed for this repository

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.