Sunday 8 December 2013

Have you ever had breakfast with Sophia Loren?

There was a very good programme on BBC4 recently, The Joy of Logic: "Documentary exploring the human quest for certainty and sound reasoning itself. Professor Dave Cliff asks, just how logical are we really and can humans stay ahead?". Highly recommended.

Never having had breakfast
with Sophia Loren,
David Moss sets out on a journey
from Plato to Bjørn Lomborg‎,
with a long pit stop
at Willard Van Orman Quine.
Why would anybody do that?
The BBC programme covers much of the same ground as the esteemed treatise on artificial intelligence penned by none other than DMossEsq himself currently basking at no. 5,298,041 in Amazon's bestsellers rankings. If even one person buys a copy for Christmas thanks to Professor Cliff, surely the book could once again shoot up into its rightful position in the high four millions.

Have you ever tried to read Gödel's Theorem? Couldn't even understand the definitions? Same here.

But it's important. In any formalised language, you can have either completeness or consistency, one or the other, but not both.

Professor Cliff reminds us of Gödel's belief that people were trying to poison him. The only person he would allow to prepare his meals was his wife. When she fell ill and went into hospital he starved to death.

Which makes the first of many important distinctions – just because a person like Kurt Gödel is highly intelligent, it doesn't follow that he's sane. Intelligence and sanity are two different things ...

... a Pandora's box which needs to be opened because the study of logic, the principles of mathematics and developments in computer science all wash up on the shores these days of artificial intelligence (AI), not to be confused with artificial insemination.

Surely the brain is like a computer. Surely the web is like a giant brain. Study neural networks and surely you're studying how the brain works.

Really?

Professor Susan Greenfield, who knows a bit about brains, is always magnificently withering about these claims of AI. Studying the computational activity of the brain is the easy bit, the bit that all the logicians and mathematicians and computer scientists just happen to have methods for. What about all the other activities that occupy the human mind? The AI bods have not a word to say about them, she points out scornfully and correctly, AI is baffled into silence.

AI is baffled into silence about human psychology and social dynamics and political governance. Not to mention aesthetic, ethical and theological judgement.

Just like the neuroscience bods who think they can study ethical decision-making by watching which bits of the brain light up during the process. Just like the econometricians who think they can analyse human motivation using nothing more than a trivial model of maximising returns. Baffled. Silence.

An AI bod may know an encyclopaediaful of facts about the web and yet still have nothing especially cogent to say about humanity. Here's professor Sir Tim-Berners Lee, for example, talking nonsense:
Armed with the information that social networks and other web giants hold about us, he said, computers will be able to "help me run my life, to guess what I need next, to guess what I should read in the morning, because it will know not only what's happening out there but also what I've read already, and also what my mood is, and who I'm meeting later on".
And that's just the genuine academics. Then there's the retail end, the burgeoning industry of the so-called "quantified self" – "There is further investment in the quantified self space as Canadian company Retrofit announces $8 million in new funding ..." – where a little more self-knowledge of bafflement might promote a little more healthy silence.

Let's get back to the engaging Professor Cliff. He opens the programme with a disarming story of his own unintelligence. He wrote a computer program years ago that emulates the activity of traders in markets – purely computational, perfectly feasible – which was instrumental in removing the human traders and replacing them with, as he says, a computer, a dog and a man. The dog's job is to protect the computer and the man's job is to feed the dog.

Why unintelligent? He gave the computer program away for free and lots of other people have made a fortune out of it. Is that so unintelligent? Discuss.

The TV programme includes a large cast of eminent talking heads. Including Professor Sir Nigel Shadbolt, professor of AI at the University of Southampton, chairman and co-founder of the Open Data Institute and chairman of the Department for Business Innovation and Skills midata programme.

He's been putting himself about a bit recently, Professor Sir Nigel.



I think that you'll find that within the open data community there is a very strong recognition of the need to protect privacy but it’s more complicated than baldly saying 'open data is not personal data.
Technical Director of the Open Data Institute
There was his 22 October 2013 double act with Stephan "embrace the change" Shakespeare, for example, giving evidence in front of the Public Administration Select Committee – there is no good reason to preserve personal privacy, not when making personal data open would benefit society.

There was his appearance in the Financial Times the day before yesterday with his Maseratis and his 30-foot sloop and his book, The Spy in the Coffee Machine: The End of Privacy as we Know It, which is going on at least one person's Christmas list.

And there was his review of the work done by the midata Innovation Lab where he said that the five prototype me-too applications produced by the lab allow us to "get to the future more quickly". Che?

Logic.

That's where we started.

You and Professor Cliff could search high and low without finding the answer to the question where is the logic in Professor Sir Nigel wasting his time and our money re-creating what the market has already provided?

----------

Updated 9-6-14

Computer passes 'Turing Test' for the first time after convincing users it is human

Updated 11.6.14

As you were:
World to Captain Cyborg on 'Turing test' stunt: You're Rumbled
Man who pretended to be a robot fooled by robot pretending to be a man

... This week, the realisation may have belatedly dawned on much of the mainstream media that a Kevin Warwick claim needs to be taken with a mine's worth of salt ...

What did it this time is Warwick's claim that the "Turing Test" - which measures ability of a machine to convincingly mimic a human while communicating with real humans in a blind test - had been passed at an event Warwick had organised and hosted. This had all the hallmarks of a Warwick stunt - you only had to look.
Updated 14.6.14

Professor Sir Nigel Shadbolt:
No, robots aren’t taking over. But the future is dazzling

Building an engaging, general and well-rounded intelligence is as far away as ever.
Updated 17.7.14
Jesus College
Oxford

Election of Next Principal

Tuesday, 15 July 2014
Members and friends of the College will be pleased to learn that the Governing Body has resolved to elect as its next Principal, Professor Sir Nigel Shadbolt, who is, at present, Professor of Artificial Intelligence at the University of Southampton. The formal election will take place in October 2014. Sir Nigel will take up the post with effect from 1st August 2015.

No comments:

Post a Comment