On Nov 22, 2005, at 2:35 AM, Gregory Maxwell wrote:
On 11/22/05, S. Woodside sbwoodside@yahoo.com wrote:
Conjecture 2. That wikipedia is sufficiently formal and complete that you could build a useful general purpose AI knowledge base using it.
On a more constructive note, If you're interested in AI taught using wikipedia... you should take a look at the state of the art in English parsers (such as [[Link Grammar]]) and how they fair on Wikipedia. Impressive that it works at all, but I don't think we need to worry about a hostile takeover of the planet by a self-aware encyclopedia any time soon.
I'm googling a little bit now for wiki-related AI projects... What stuck in my mind for conjecture 2 was the Loebner Prize, where you create these chatbots to try to win a "Turing Test" style competition. One of the biggest challenges they have is to come up with a good knowledge base, and since Cyc isn't free, well, that's a challenge. Could you build a good Loebner competitor using wiki? I have no clue. :-) But if someone put a gun to my head that's probably the track I'd go down.
It's not like I think I'm some kind of scientific genius but I did run these conjectures by a friend of mine who I think is a genius and he seemed sufficiently interested that I thought I'd spam the world with them ;-)
--simon