AI ALIGNMENT FORUM
AF

HomeLibraryQuestionsAll Posts
About
HomeLibraryQuestionsAll Posts

All Posts

Sorted by Magic (New & Upvoted)
Timeframe:All TimeDailyWeeklyMonthlyYearly
Sorted by:Magic (New & Upvoted)Recent CommentsNewOldTop
Filtered by:All PostsQuestionsMeta
Show Low Karma

Monday, October 26th 2020
Mon, Oct 26th 2020

Sunday, October 25th 2020
Sun, Oct 25th 2020

Saturday, October 24th 2020
Sat, Oct 24th 2020

No posts for October 24th 2020

Friday, October 23rd 2020
Fri, Oct 23rd 2020

Thursday, October 22nd 2020
Thu, Oct 22nd 2020

Wednesday, October 21st 2020
Wed, Oct 21st 2020

Shortform
1Alex Turner6dFrom unpublished work.The answer to this seems obvious in isolation: shaping helps with credit assignment, rescaling doesn't (and might complicate certain methods in the advantage vs Q-value way). But I feel like maybe there's an important interaction here that could inform a mathematical theory of how a reward signal guides learners through model space?

Tuesday, October 20th 2020
Tue, Oct 20th 2020

Monday, October 19th 2020
Mon, Oct 19th 2020

No posts for October 19th 2020

Sunday, October 18th 2020
Sun, Oct 18th 2020

No posts for October 18th 2020

Saturday, October 17th 2020
Sat, Oct 17th 2020

No posts for October 17th 2020

Load More Days