Tag Archives: artificial-intelligence

Ian McDonald on our digital doppelgangers

DSC_0024The BBC is running an essay by Ian McDonald, author of Brasyl and River of Gods (and many more sf novels). Despite being an deliberate laggard on social network and metaverse platforms himself, McDonald suggests that the science fictional trope of the uploaded human consciousness is already becoming true by degrees:

Our You2s will ever more closely resemble us, and become more and more intelligent as they make linkages between the information we placed there. They’ll take decisions without our interference -and they’ll increasingly talk to each other. It’s no coincidence that the net is shaped like a society.

Perhaps there will never be a single moment when computers become aware. Maybe it will be a slow waking and making sense of that blur of information, like a baby makes sense of the colour patches and patterned sounds into objects and words.

Why should artificial intelligences – our You2s – take any less time to grow up than us?

Artificial intelligences make regular appearances in McDonald’s fiction – and he’s a writer I recommend without hesitation to any science fiction reader – though here it’s almost as if he’s conceding that a kind of ‘soft takeoff’ Singularity is already in its early stages in the real world.

Being a good science fiction writer, though, he’s considering the implications of the future:

What we’ll have is a copy of a personality in a box. It’ll be you in every detail that makes the meat-you you. You2. Only it’s technically immortal as long as the hardware keeps running and is regularly updated. This sounds great, until you realise that the original you still goes down that dark valley from which there is no return…

Quite a synchronous topic, really, given the recent flare-up of Singularitary debates. [Hat tip to Ian Sales; image by your humble correspondent.]

Reverse engineering the brain

thinkResearchers describe how it might one day be possible to simulate large parts of the human cortex on a computer, and how this could lead to functional human equivalent AI:

Software simulation of the human brain is just one half the solution. The other is to create a new chip design that will mimic the neuron and synaptic structure of the brain.

That’s where Kwabena Boahen, associate professor of bioengineering at Stanford University, hopes to help. Boahen, along with other Stanford professors, has been working on implementing neural architectures in silicon.

One of the main challenges to building this system in hardware, explains Boahen, is that each neuron connects to others through 8,000 synapses. It takes about 20 transistors to implement a synapse, so building the silicon equivalent of 220 trillion synapses is a tall order, indeed.

This is a different approach to more traditional AI research that has been going on for decades: instead of trying to write artificially intelligent computer programs using knowledge representation or commonsense knowledge representation now researchers are concentrating on reverse-engineering the only extant example of general intelligence we have.

[at Wired][image from bschmove on flickr]

Spam-trap Turing tests train smarter software

Email and comment spam is one of those constant low-grade annoyances that simply becomes part of the furniture if you spend a lot of time on the ‘net, as are the CAPTCHA puzzles you have to take to prove you’re a human. [image from Wikimedia Commons]

Signs are that they won’t be much use much longer, though; a UK researcher has been using the ‘twisted letters’ type of CAPTCHA to train his visual recognition algorithms, while a chap at Palo Alto has a program that can correctly identify cats and dogs 83% of the time – which, lets face it, is probably a better success rate than the average YouTube user can manage.

Sadly, training algorithms against Turing test spam-traps is no more likely to produce a recognisably intelligent piece of software than the Loebner Artificial Intelligence Prize is. But maybe one day we’ll be able to combine all the pieces… if they don’t beat us to it and combine themselves, of course. 😉

Loebner Prize winner doesn’t believe in Turing Test anyway

Yesterday saw Reading University here in the UK playing host to the annual Loebner Artificial Intelligence Prize event – a contest based around Alan Turing’s famous benchmark for artificial intelligence that can really think, namely whether or not it can successfully imitate human communications.

The bronze medal (for fooling a quarter of the judges) went to Elbot, a chat-bot program created by Fred Roberts, but Roberts himself seems to be not so impressed by Turing’s theory:

“I don’t think it’s anything like thought,” he said of Elbot’s conversational prowess. “If you know a magic trick, you know how it’s done, it’s not magic anymore. Sorry to be so pessimistic.”

With the caveat that I have no expertise in cognition or expert systems, I’m inclined to agree with him. [via The Guardian]

Computers react before humans have a clue

800px-Japanese_car_accident SF stories involving artificial intelligences often play up the fact that a computer-based intelligence would find human thought processes glacially slow in comparison to its own.

But you don’t have to dip into speculation about the future of computing to see that. The Australian newspaper The Age points out that although “survivors of serious car crashes often say time appears to slow down in the moments around the impact and that they can recall the event in extraordinary detail,” in reality, “the crash is often over before the human brain has registered the incident, and it’s only by later replaying it in their minds that crash victims achieve such vivid recollections.” (Via Instapundit.)

Accompanying the story is this anatomy of a crash which makes their point:

All over in the blink of an eye

This is a reconstruction of a crash involving a stationary Ford Falcon XT sedan being struck in the driver’s door by another vehicle travelling at 50 km/h.

One millisecond equals 1/1000th of a second.

0 milliseconds – An external object touches the driver’s door.

1 ms – The car’s door pressure sensor detects a pressure wave.

2 ms – An acceleration sensor in the C-pillar behind the rear door also detects a crash event.

2.5 ms – A sensor in the car’s centre detects crash vibrations.

5 ms – Car’s crash computer checks for insignificant crash events, such as a shopping trolley impact or incidental contact. It is still working out the severity of the crash. Door intrusion structure begins to absorb energy.

6.5 ms – Door pressure sensor registers peak pressures.

7 ms – Crash computer confirms a serious crash and calculates its actions.

8 ms – Computer sends a “fire” signal to side airbag. Meanwhile, B-pillar begins to crumple inwards and energy begins to transfer into cross-car load path beneath the occupant.

8.5 ms – Side airbag system fires.

15 ms – Roof begins to absorb part of the impact. Airbag bursts through seat foam and begins to fill.

17 ms – Cross-car load path and structure under rear seat reach maximum load.
Airbag covers occupant’s chest and begins to push the shoulder away from impact zone.

20 ms – Door and B-pillar begin to push on front seat. Airbag begins to push occupant’s chest away from the impact.

27 ms – Impact velocity has halved from 50 km/h to 23.5 km/h. A “pusher block” in the seat moves occupant’s pelvis away from impact zone. Airbag starts controlled deflation.

30 ms – The Falcon has absorbed all crash energy. Airbag remains in place. For a brief moment, occupant experiences maximum force equal to 12 times the force of gravity.

45 ms – Occupant and airbag move together with deforming side structure.

50 ms – Crash computer unlocks car’s doors. Passenger safety cell begins to rebound, pushing doors away from occupant.

70 ms – Airbag continues to deflate. Occupant moves back towards middle of car.
Engineers classify crash as “complete”.

150-300 ms – Occupant becomes aware of collision.

So if the Singularity every arrives, does this mean it will all be over before humans even notice it’s begun?

(Image: Wikimedia Commons.)

[tags]automobiles, safety, computers, artificial intelligence[/tags]