“the algorithm I wrote”—where might I find this algorithm?
etc
General practice is to have links from new terms.
I see that most of your postings seem to have similar issues. Even in pure math papers they tend to put references so the reader has some chance to work out what you are talking about.
This was originally posted to the AI Alignment Forum, where researchers are typically engaged in an advanced technical discussion. AI Alignment Forum posts assume these concepts; they aren’t new terms. This is similar to how pure math papers don’t explain what groups are, what homomorphisms are, etc.
If you want to get up to speed so that you can understand and contribute, I suggest some googling (here’s the wiki page and LW wiki page for AIXI). There’s also a lot of historical discussion of AIXI on LW that you can read.
(I think we will update the posts that are from AIAF to include info at the top about them being from the AIAF.)
Yes, UDT means updateless decision theory, “the policy” is used as a placeholder for “whatever policy the agent ends up picking”, much like a variable in an equation, and “the algorithm I wrote” is still unpublished because there were too many things wrong with it for me to be comfortable putting it up, as I can’t even show it has any nice properties in particular. Although now that you mention it, I probably should put it up so future posts about what’s wrong with it have a well-specified target to shoot holes in. >_>
This post would be much helped by some explanation and context.
AIXI—I happen to know what this is but maybe not everyone does
UDT—Maybe Updateless Decision Theory? Maybe something else?
https://en.wikipedia.org/wiki/UDT
“the policy”—what policy?
“the algorithm I wrote”—where might I find this algorithm?
etc
General practice is to have links from new terms.
I see that most of your postings seem to have similar issues. Even in pure math papers they tend to put references so the reader has some chance to work out what you are talking about.
This was originally posted to the AI Alignment Forum, where researchers are typically engaged in an advanced technical discussion. AI Alignment Forum posts assume these concepts; they aren’t new terms. This is similar to how pure math papers don’t explain what groups are, what homomorphisms are, etc.
If you want to get up to speed so that you can understand and contribute, I suggest some googling (here’s the wiki page and LW wiki page for AIXI). There’s also a lot of historical discussion of AIXI on LW that you can read.
(I think we will update the posts that are from AIAF to include info at the top about them being from the AIAF.)
Yes, UDT means updateless decision theory, “the policy” is used as a placeholder for “whatever policy the agent ends up picking”, much like a variable in an equation, and “the algorithm I wrote” is still unpublished because there were too many things wrong with it for me to be comfortable putting it up, as I can’t even show it has any nice properties in particular. Although now that you mention it, I probably should put it up so future posts about what’s wrong with it have a well-specified target to shoot holes in. >_>