Showing posts with label AI. Show all posts
Showing posts with label AI. Show all posts

21.7.09

accolades

turns out i won two prizes in CS this year, one for first-year AI (which i don't think i shoulda got since i'm in second year, but they really liked my article on using AI techniques in security, and i got the best grades in the exam, natch) and the other the O'Reilly Prize in Computing Science. they're giving me a mystery parcel that i gotta go pick up, and apparently at the welcome session this year the prizes are good for £150 worth of books. (the mystery parcel is taunting my curiosity by waiting in a Department office until Thursday.)

i'm gonna be able to afford textbooks!

L

PS. who knows, maybe i will make top of the class in something next year.

30.10.08

i'm a bastard

i am. i and a few other people made a guy storm out of our Grand Challenges of AI tutorial today just by questioning him.

the tutorials are pretty simple for GCAI, which is an introductory, non-technical artificial intelligence primer that i take for giggles (can't get any credit for it, for various reasons.) you get a source book of articles, you bring it to each tutorial, and each time, you read through one of them and a sort of informal town-hall debate arises around it. today it was Turing's seminal paper, Computing Machinery and Intelligence, from way back in 1950. it's basically the original introduction to the field of AI, where Turing defines intelligence, presents the concept of machine intelligence, and demos the common arguments against the possibility of deep AI. he then rebukes each one, apart from the last, which is (hilariously) the 'argument from telepathy'; at the time, it hadn't yet been discredited, and Turing was a firm believer. it shows in the paper, with a genius like Turing asserting the existence of evidence for telekinesis and telepathy which he doesn't provide and suggesting the necessity of 'telepathy-proof rooms' for testing AI in.

that wasn't the argument we had problems with, however. Turing's first cited objection to AI is religious, the 'Theological Objection'. people apparently didn't like AI, back in the day, because of their ridiculous conviction that only people could think because only people had souls. of course, a machine could never have a soul, so how could it possibly ever be intelligent?

i pointed out that this shouldn't even have been included in a scientific paper in the first place. it's just pandering to the deluded, pretending that their silly beliefs are legitimate enough to warrant consideration in a serious scientific journal, but it was the norm in Turing's time to pamper these people. i get that, but, i explain, you can't argue from God unless you've previously proved God, and you can't make arguments involving souls unless you've demonstrated their existence already. which nobody has.

crazy two rows down from me pipes up. "Just because it involves God is no reason to automatically say it's bullshit," he tells me sternly.
"It is," i throw back. crazy goes on to explain that we have souls, animals don't have souls, machines don't have souls, and that therefore we're fundamentally different from either one and the only intelligences possible.

why is this guy studying AI? it being pretty harsh to say that, i stick with asking him for evidence of his batshit claims, and with giving an explanation of neural nets and how the ones we have are tiny arrays of perceptrons compared to our trillions of neurons. we're not anywhere close to creating an analogy of the human brain, but if we did get that many perceptrons, there's no reason to write off the possibility that they might approximate a sentient intelligence. "If we can build nets as big as our brains," i ask him, "then how can you be so sure that they won't come close to intelligence?" somewhere in this i used the phrase 'biological machines' to describe us. it turns out this was a big fucking mistake. crazy didn't like that one bit.

"We're pointless then," he says, over and over. "We're just pointless then aren't we."
"We're not pointless. That doesn't follow at all," i cut him off when he starts to annoy me. "Knowing how our brains work doesn't mean we're devoid of all meaning. Or are you trying to tell me we need a soul to have meaningful lives?"

at this, crazy finally has enough. he turns round, points at me and the people who've been agreeing with me like we were witches, and goes, "I don't like this, this is just fucking shit. Fuck this. You're all getting so overemotional. Fuck this. Fuck," and he shoves his way out of the lecture theatre and stomps away down the corridor, leaving me feeling like a total asshole and my comrades giggling their asses off at this guy.

"We're getting emotional?" snorts the guy just in front of me as the crazy's footsteps die away. i still feel like a shit about it, but i can't just sit there in those sessions and let people spew bullshit unchallenged. i guess some people just can't cope with being asked to actually prove what they're claiming is true.