AI / General Chatting / Forums

Forums

4hv.org :: Forums :: General Chatting

« Previous topic | Next topic »

previous 1 2 3 next

Move Thread

LAN_403

Conundrum

Tue Nov 04 2014, 06:25PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

I seem to recall reading that the Turing test has in fact not been passed, as the subset of imitation was too narrow.
IIRC 75% of judges have to be convinced that the machine is in fact a human being, and at leat 25% that the human being is in fact the machine.
This has not happened yet, although the software is indeed a significant step in the right direction.

Dr. Slack

Tue Nov 04 2014, 08:51PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

The Turing test is something of a red herring. The equivalent for mechanical engineering would be sticking your toes through a hole in the screen, and trying to work out whether they were being smashed by an engine-driven hammer, or a person with a rock.

If you sat down to watch a film, and the opening 20 minutes consisted of people saying 'we need to communicate in natural language, so we need to get the machines to learn to understand people, and by the way, let's network this all together with machines that control our vital supplies, and machines that can make new machines', you would be saying 'ooo, ooo, I can guess what happens in the second half of the film!'

hen918

Tue Nov 04 2014, 09:01PM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

I like that analogy! Turing never did intend this "test" to be the ultimate benchmark for AI.

Conundrum

Wed Nov 05 2014, 07:43AM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

Heh.. same here.
My research suggests that 21/2/18 is plausible only if RTSC is discovered in early 2015 which is the enabling technology that allows molecular level interconnects between <18nm neural processors to form.
These would be somewhat self repairing and thanks to oxygen annealing be able to dynamically form and deform at will in a broadly similar way to memristors.

Significant advantages of this system are: no resistive losses within the material, presence of a Mott insulator makes the neural links able to carry more data without crosstalk and the nature of superconductors means that the normally overheating neural processors would all be at the same temperature which allows a relatively simple cooling system to be used.

Uspring

Wed Nov 05 2014, 10:01AM

Registered Member #3988 Joined: Thu Jul 07 2011, 03:25PM
Location:
Posts: 711

Dr. Slack wrote:

The Turing test is something of a red herring. The equivalent for mechanical engineering would be sticking your toes through a hole in the screen, and trying to work out whether they were being smashed by an engine-driven hammer, or a person with a rock.

If it would be _very_ difficult to build a engine-driven hammer, this would indeed be an interesting test.

Ash Small

Wed Nov 05 2014, 11:37AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Uspring wrote ...

If it would be _very_ difficult to build a engine-driven hammer, this would indeed be an interesting test.

Engine driven hammers have been around for centuries:

Conundrum

Thu Nov 06 2014, 07:55AM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

Also worth mentioning, we don't fully understand the mechanism of HTSC in the first place.
If it turns out that the CDW/SC mechanism is indeed correct then it could be that ETSC is actually easy once you have the correct formula.

On an aside, how likely is it that superconductive materials occur in nature given that we have found pnictide materials that work up to 109K?

EDIT:

DÃ©dÃ©!

Sat Nov 29 2014, 01:48PM

Registered Member #4932 Joined: Thu May 17 2012, 01:42PM
Location:
Posts: 59

I haven't read everything yet, but i found this was interesting.
http://www.research.ibm.com/cognitive-computing/neurosynaptic-chips.shtml#fbid=XCjEQmUBDqs

Conundrum

Sun Dec 14 2014, 03:42PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4062

BTW, duff link.
Also it seems that Y123 (aya YBCO) can apparently superconduct near room temperature if irradiated with an infrared laser at a specific wavelength at least on the surface of the crystal.

Patrick

Mon Dec 15 2014, 12:34AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Dr. Slack wrote ...

I love Elon Musk's question, are we just a biological boot loader for the digital intelligence to come?

Self replicating gray or green goop will kill us all, long before any of these questions can or need an answer.
Humanity is obsolete, Lithium and Prozac helps to ease the scariness of this realization.

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob