Topic-Seg-Label icon indicating copy to clipboard operation
Topic-Seg-Label copied to clipboard

Published 20 hours ago • verified publisher icon truthless11

Delayed Reward

Open waretupper opened this issue 4 years ago • 0 comments

The delayed reward defined by def HSMI(batch_obs, action_list): line 12 of pn_tool seems to only contain the element that encourages lower similarity between adjacent segments. Unless I am mistaken, compared the equation presented in the paper it seems the element encouraging higher similarity within a segment is missing.

waretupper avatar Jun 11 '20 12:06 waretupper