Comments on: Thoughts on RL with a changing action/state space /stor-i-student-sites/jack-trainer/thoughts-on-rl-with-a-changing-action-state-space/?utm_source=rss&utm_medium=rss&utm_campaign=thoughts-on-rl-with-a-changing-action-state-space My Personal Blog Wed, 06 Mar 2024 12:08:17 +0000 hourly 1 https://wordpress.org/?v=7.0.3 By: Vitor Martins /stor-i-student-sites/jack-trainer/thoughts-on-rl-with-a-changing-action-state-space/#comment-4324 Wed, 06 Mar 2024 12:08:17 +0000 /stor-i-student-sites/jack-trainer/?p=1227#comment-4324 But if the action space and state space was shaped only to the leggal actions, i.e. the Neural Network is modeled accounting that, the training time won’t be smaller ?

]]> By: Heng /stor-i-student-sites/jack-trainer/thoughts-on-rl-with-a-changing-action-state-space/#comment-2628 Tue, 26 Sep 2023 14:00:15 +0000 /stor-i-student-sites/jack-trainer/?p=1227#comment-2628 Hi, do you have some examples or papers about your mentioned first method?

]]> By: feifei /stor-i-student-sites/jack-trainer/thoughts-on-rl-with-a-changing-action-state-space/#comment-2617 Sun, 24 Sep 2023 03:49:23 +0000 /stor-i-student-sites/jack-trainer/?p=1227#comment-2617 your article is great! Currently I am exploring the ideas about how to resolve non-fixed sized action space and non-fixed sized and complicated state space. I will need to check on the pointer network.

Thanks.

]]>