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

Entry: The Cyc - A959123
Author: raymondo - U202611

Hey I left out all references to Deep Thought. Do I get any brownie points?

this entry is a bit of a pitty, especially when you consider that the author has heard of the cyc project of doug lenat and chose instead to use it as inspiration for an unoriginal peice of fiction.

if he could be bothered to turn this into a serious review about the cyc project, what it is, and why, then it might be worthy of being in the edited guide, but at the moment, it just needs stuffing back onto their userspace as a journal entry.

http://wombat.doc.ic.ac.uk/foldoc/foldoc.cgi?Cyc covers it rather well.

Hi Raymondo, hi Xyroth,

just to say that i read this and whilst i can sympathise that it is not exactly EG-style, it seems a shame to be so harsh on it just for that reason.

The content is new to h2g2, the style is good, though not EG, and the author is relatively new and still finding their feet...

Good stuff, Ray! or just: 'Whay! Ray!'



Tis true, what X says, that this will not make it into the EG in this particular form, but you *do* get brownie points for content and style!

keep on truckin'
spiff

I seem to recall reading about the event in WIRED magezine during an interview with it's project manager or creater. I will lookup the reference and get back to you.

Raymondo, Xyroth,

Why has Xyroth labelled this as fiction? Did the incident described take place or did it not take place?

I would object to the interpretation of the incident as "self-awareness". I'd rewrite that bit to give self-awareness as a possible explanation rather than THE explanation.

I think the "incident" takes up too much of the entry. We need an explanation of the problems of computers and common sense. Some more examples of a computer "knowing" facts but not understanding them. And then an explanation of how the cyc might solve this.

"why has xyroth labeled it as fiction?" - because he didn't read it properly and under those circumstances that is how it came across.

sorry about that.

however the "incident" is the thing that gives this impression, and as others have said, it takes up much too much of the article.

It would probably be better if it gave a more thorough explanation of what the cyc project is, and how one of the big names complained about nobody doing this work, so lenat took it up to take over from his eurisko program.

the "wired" write-up might be http://www.wired.com/wired/archive/5.01/ffhal_pr.html which took place in 1997 or http://www.wired.com/wired/archive/6.11/wired25.html?pg=11 in 1998.

he might want to link to the main site for cycorp who run the cyc project at http://www.cyc.com/

I have also found while exploring the cycorp website the opencyc.org project, where people like us can get our hands on the opensource version of the program and some of it's knowledgebase.

its at http://www.opencyc.org/

Hey, Ray, still planning on working on this one?


Mikey

yes, was busy on A998841 - brains the size of a solar system...

the "big name" I mantion a couple of posts ago was marvin minsky.

he criticised the then current expert systems as being toy programs, which only work in small knowledge domains, and were no good at discovery ofnew stuff and learning.

lenat built the programs AM (Amatuer Mathematician) and Euriskoto prove that wrong, and the success of these programs allowed him to get funding for the cyc project, to encode "common knowledge" into a computer.

Hi there!

Are you going to find time to work on this or would you prefer to remove it from Peer Review?

Whoami?
h2g2 Scout

He had made me a bit of a promise a while back about clearing all these out (or, alternatively, getting them in gear) as soon as he came back from his Jamaican vacation, but that seems to have fallen through.

aggggggh!
caught again, curses
I absolutely promise to work on all outstanding entries this weekend and maybe even Thursday

I could take an oath not to play Starcraft until they are complete, but I would just break it....

oops, I just spotted that I got the name of "euroskno" wrong and called it "Euriskoto" by mistake.

the stated aim of lenat was to leave the euroskno program in limbo, while he built up enough "common knowledge" in the cyc database for it to be let loose on.

it would then be let loose on the database and rule base, and see what it could find from there.

interestingly, about a quarter of the stuff cyc now learns, it learns just by reading it, and having enough common knowledge to be able to find the context for the new stuff.

looking forward to reading the rewrite.