h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

Entry: Machines and Intelligence - A582284
Author: xyroth (getting swamped in the backlog) - U149792

well, this is the last of the first batch from the intelligence project.

please ask about anything you don't understand.

Again needs combining with some of the others.

fine mckay, instead of just saying "combine them", and on another thread "more work required", how about making actual usefull comments as to how they might be combined, or which bits need more work.

I'm all in favour of criticism and feedback, that is why I put the entries here, but the emphasis is on "constructive".

You've got to remember, I'm reading all these seperate pieces on Intelligence, without knowing whats in each one until I've read it. Not greatly helped by my having been on 3 weeks holiday and having an overflowing conversation list.

I was adding a bit to each to show I'd read them - overall (on what I've read so far) I think firstly you need a piece defining what you are calling intelligence and a bit about different forms of intelligence.

Then a piece about measuring intelligence by the different methods - (or one on each if there's sufficient.)

Then a piece on exceptions to what you've written (idiot savants, autistics, etc.)

Then a piece tying it all together, with a future directions bit in it.

This is just a view of someone who has not been involved in what has obviously been considerable ammounts of previous work, and I know the time constraints you're up against, but any piece can come into Peer Review at any time, and I think that would be a better way to take it.

Intelligence is fascinating, in that we all know what we mean by it, but none of us can define it, and we all use different measures for it.

Good luck with this, and I'm happy to help where I can, but its a big baby.

about that, but I have seen too many people in the history of this project come along and say "it's crap" or a similarly unhelpfull comment in multiple threads, and then say no more.

and you are slightly wrong about "you can put entries in peer review anytime". specifically, when there are multiple entries, they have been traditionally moved to a university project, but if wedon't get a couple of entries through, then it won't get the credit.

so there are time constraints on some of this.

the persistant problem has been lack of feedback. although there was over 90K of stuff in the original infamous intelligence thread about the subject, it has proved a real pain to find the rightplaces to split it up.

in the end I was forced to split it into a seperate entry for each key idea. while I am not against the idea of merging some of the entries, it needs to be done in such a way as to make it easy to split them again when extra information starts to overload them.

please be generous in your amount of comments.

Hi Xyroth!

You should either submit these entries on Intelligence to Peer Review one at a time, waiting for each to be accepted before submitting the next, or you should combine them all into one big one. That way, you will see all the problems there are in each entry, rather than jumping from one to another.

As a self contained entry on Machines and Intelligence, this one needs a lot of work. The first section is really "Issues raised by Machine Intelligence", while the second section is a general introduction. They should be switched so that the introduction comes first. The section called Summary is not a summary at all, it is just an additional comment.

right, I have swapped the first two around and retitled the closing comment.

I have tried the "one big entry" model, and the subject is just too large for it.

I have also tried working on one at a time, with the result that it introduced extra redundancy across the entries which has been criticised in some of the other threads.

I tried working the project seperately, with the result it was just impossible to get feedback on the entries.

I have not disputed that the entries need work, only that I have gone as far with some of them as I can without feedback.having just received your feedback, I have made the suggested changes.

any more suggestions?

That first header isn't needed - the title forms the first header.

Hi Xyroth,

Interesting reading

This post is little more than a bookmark - I'll be back, cos the subject interests me, but whilst I' here, one point...

There's nothing (that I can see) about intelligence being a (or _the_) product of a machine learning by it's mistakes, which is what I thought AI was about...


Trivial typos:
<< computer languages like prolog and lisp >> ...Prolog and Lisp...

Xyroth? Are you still here and working on this?