Hal is born

Post #6,043 by DonRichards 8/21/01 10:09:49 AM Reply	Hal is born A team of scientists in Israel have created what just may be the first A.I. able to pass the Turing test. Taking Turing's advice, instead of trying to program a system with the thousands upon thousands of facts needed to emulate adult thinking, they started with a "child" that knew only the alphabet and has a preference for rewards and an aversion to punishement. Already able to fool child development experts into thinking they are conversing with a toddler, the system is estimated to need 10 years of learning before it will be able to pass the Turing test. BTW, they did indeed name the system "Hal" in honor of the Kubrick/Clarke classic. [link\|http://www.cnn.com/2001/TECH/industry/08/20/computer.hal.reut/index.html\|[link\|http://www.cnn.com/2001/TECH/industry/08/20/computer.hal.reut/index.html\|http://www.cnn.com/...t/index.html]] "When it crosses my mind to do something, I don't ask why, I ask why not. And usually there's no reason not to, so I just go ahead. It's given me the strangest collection of hats"
Post #6,124 by tseliot 8/21/01 3:41:55 PM Reply	Halfway there. I was going to make this an independent topic, but you had to go and bring up AI at the same time, din'cha? ;) I had the most amazing conversation last night at a Robot Wars premiere party: I spent at least a half-hour talking to one of the staff on the [link\|http://www.nsi.edu/nomad\|NOMAD] project. This is the FIRST instance of an AI I can support theoretically, because it has a body as well as a brain. The website unfortunately does not go into all the details, but the breadth of neurobiological simulation they are continually adding to this baby is staggering. Now if I can just get them to give it a heart and lungs... :) Gonna see if I can volunteer for them. That's her, officer! That's the woman that programmed me for evil!
Post #6,435 by neelk 8/23/01 9:15:42 AM Reply	You should also look at the Cog project, then... [link\|http://www.ai.mit.edu/projects/humanoid-robotics-group/\|http://www.ai.mit.e...otics-group/] Side note: I once took a class with Rod Brooks. Having met him, you can understand how he is able to decide the entire AI research community is wrong, and you can also understand why he's able to go out and invent robots that prove it. He really is that cynical, and that smart. :) Quote: "The central idea that I've been playing with for the last 12-15 years is that what we are and what biological systems are. It's not what's in the head, it's in their interaction with the world. You can't view it as the head, and the body hanging off the head, being directed by the brain, and the world being something else out there. It's a complete system, coupled together."
Post #6,445 by deSitter 8/23/01 11:51:02 AM Reply	Fascinating The quote was the starting point of Jung and Pauli's joint research into the phenomenon of synchronicity. For myself, I express the same idea by stating that memories are not stored in the brain, as RAM, rather they are enfolded into the fabric of reality, and remembering is then an act of unfolding. I was watching a [link\|http://www.conservation.state.mo.us/nathis/arthopo/mospider/kinds.htm#anchor866508\|garden spider] build her magnificent web yesterday. She moved with incredible speed and certainly, using her legs to test the tension in the web as she layed out orbital threads. Since a web has to conform to its environment, compromises are needed at certain points - the big gal would stop, seemed to be thinking, then she would set off to the other side with a new spoke and start laying down orbits again. It was amazing to watch! There is no way this entire complex procedure could be stored in its tiny brain. Somehow spider existence is enfolded into reality to the tune a few hundred million years, and all that is necessary to induce unfoldment of web-building is an empty stomach.
Post #6,494 by tseliot 8/23/01 2:21:45 PM Reply	Thanks! Cursory view looks promising... That's her, officer! That's the woman that programmed me for evil!
Post #6,228 by wharris2 8/22/01 2:50:41 AM Reply	Hal, shmal Unfortunately it doesn't seem like "hal" is available to the general publc. Hmmmmmmm. This naturally gives rise to certain doubts. There has been a lot of research about chess playing and AI. Can Hal learn how to play chess? Oops, sorry, we don't have any access to Hal. At the moment, I don't care much how well Hal plays chess, just that he can learn how to play chess. If he can't, it instantly disproves the supposed "AI" bullshit. If it can, it is an impressive accomplishment even if it doesn't entirely fulfill the Turing test.. French Zombies are zapping me with lasers!
Post #6,256 by DonRichards 8/22/01 9:09:58 AM Reply	Why so hostile? It was just a freakin news story. No need to get your panties in a knot. Wait till the ten years is up, then see if it can learn chess. As far as the general public having access, how many other scientific endeavors allow access by the general public? Doesn't seem to unreasonable to me. "When it crosses my mind to do something, I don't ask why, I ask why not. And usually there's no reason not to, so I just go ahead. It's given me the strangest collection of hats"
Post #6,275 by wharris2 8/22/01 9:57:27 AM Reply	Just the hype "Israeli computer passes Turing test".... now to me, that's outrageous hype without further supporting detail or other tests. French Zombies are zapping me with lasers!
Post #6,297 by DonRichards 8/22/01 11:15:32 AM Reply	But it doesn't say that It specifically points out that it may pass a Turing test after 10 years of "learning". I thought the article was remarkably hype free for a story regarding AI. I do agree that most "news" about AI is a hype fest. I just didn't think this was one. "When it crosses my mind to do something, I don't ask why, I ask why not. And usually there's no reason not to, so I just go ahead. It's given me the strangest collection of hats"
Post #6,268 by ChrisR 8/22/01 9:44:10 AM Reply	Validity of the Turing Test? I have my doubts about whether the Turing test really should be the litmus test for determining when a machine attains intelligence. The first problem with the test is that it says nothing about what the machine thinks about itself - or even whether it has a concept of self. Second problem is that it doesn't rule out the question of deception. The Turing test doesn't so much measure the intelligence of the machine as it does the gullibility of the human participants. Can they be fooled into thinking the machine is a fellow human? Anyhow, list me as one who thinks that the Turing test for intelligence tells us more about the nature of the human participants than it does the machine.
Post #6,298 by DonRichards 8/22/01 11:17:00 AM Reply	Good points But until something better is devised, I don't see many other options. "When it crosses my mind to do something, I don't ask why, I ask why not. And usually there's no reason not to, so I just go ahead. It's given me the strangest collection of hats"
Post #6,363 by ben_tilly 8/22/01 4:17:48 PM Reply	The best argument against the Turing Test is Eliza
Post #6,450 by drewk 8/23/01 12:05:06 PM Reply	No, the best argument against the Turing Test is Usenet Turing test is only a measure of linguistic compatibility. We can learn new words, but that's not the same as a new language. We can learn new grammar, but that's not the same as context. With sufficiently different context, anyone could fail a Turing test. Test an Aborigine via a text-only interface and I doubt he would "pass." This is my sig. There are many like it, but this one is mine.
Post #6,491 by deSitter 8/23/01 2:11:31 PM Reply	Don't You Need Numbers? The Hopi Indians had only three numbers - one, two, and many. Somehow this entire Turing thing seems hokey to me. In any case reality obeys the principle of complementarity and the superposition principle, and that can't be modeled as 1s and 0s since there is an inherently statistical element.
Post #6,592 by ben_tilly 8/24/01 10:25:50 AM Reply	Well purl passed the IRC test a long time ago Infobots have for several years been good enough to fool people into believing that they were just helped out by a real person. But yes, maintaining a coherent Usenet personality that looks somewhat normal would take some real work. Either that or it would act like Ashton. Speaking of which, Ashton, are you real...? :-) Cheers, Ben
Post #6,612 by drewk 8/24/01 12:43:41 PM Reply	How 'bout the magical Frenchman? This is my sig. There are many like it, but this one is mine.
Post #6,865 by Ashton 8/27/01 5:16:44 PM Reply	unReal.. but real-enough for Govt. Work. merely the facade Precisely as one ranks the other artifacts within the daily illusion we generate, on 'awakening'. And call those 'real'. ('You', of course would disappear were 'I' to disappear) Now as to the Real, I could only say tha Error 42: Undefined for this set. Topology error. Syntax error. Script error. Compile error. Program error. Execution error. Life systems terminated. SYSTEM HALTED
Post #6,368 by tablizer 8/22/01 4:24:38 PM Reply	Re: Fooling the Turing Test First of all, emulating the speach of an 18 month-old is not that hard. But, extrapolating that to a 5-year-old and then further is a bit too presumptuious. I can build an emulation of a 9 month child, although my Puke-A-Tron still has a few glitches in it. Did Turing specify what kind of human was to be emulated? The person on the other side could possibly mistake some of the current chat bots for skiztofranic humans. However, they will never be able to emulate my bad spelling. That should be the real test. ________________ oop.ismad.com
Post #7,274 by mhuber 8/30/01 5:27:33 PM Reply	Yep, Turing specified A man pretending to be a woman. Which was the success criteria, as well. The computer had to fool people as often as the man. Alan Turing's test, the way he wrote it, is pretty hard. The variations are much easier. I had an entirely functional emulator named Katy.exe that was indistinguishable from a person suffering from catatonia. And I didn't even write it on purpose, I renamed another program that happened to work that way due to something stupid I did with a pointer. White guys in suits know best - Pat McCurdy

Welcome to IWETHEY!