Showing posts with label Alan Turing. Show all posts
Showing posts with label Alan Turing. Show all posts

Saturday, January 7, 2017

Run, Turing, Run!

Maybe Forrest Gump invented the principles of
modern computing and artificial intelligence...

Tuesday, December 11, 2012

Magicians reveal what the world's all about

For several years, I've been fascinated by the popular books of three physicists: Brian Greene, David Deutsch and Lawrence Krauss.

Funnily enough, although the three of them are writing about the same general subject—our state-of-the-art understanding of the nature of the universe—they rarely, if ever, get around to handling the same questions in comparable, if not similar, fashions. Moreover, in their latest books, they hardly even refer to one another's work.

It's easy to understand superficially why Greene, Deutsch and Krauss don't seem to have a lot to say to one another. Greene has been reputed for a long time as an adept of string theory, and there's no reason to imagine that the other two physicists are particularly keen on this theory. Earlier this year, Krauss became widely known through his presentation of an esoteric explanation of how the ultimate "free lunch"—obtaining something from nothingness—is a perfectly plausible phenomenon at a cosmic level.

As for the 59-year-old Oxfordian David Deutsch, he comes through to me as the most philosophical member of the trio. Indeed, he offers us a multiverse view of existence that is totally amazing. As in his first book, The Fabric of Reality, Deutsch pursues in The Beginning of Infinity his quest for a Theory of Everything inspired by the work of a somewhat heteroclite foursome: Karl Popper (epistemology), Hugh Everett (multiverse theory), Alan Turing (computation) and Richard Dawkins (evolution). Indeed, between the Popperian explanations of knowledge, the connotations of quantum theory leading to the existence of multiple universes, the vast theories of classical computing put forward by Turing (which are no doubt sufficient to handle, not only the DNA computer responsible for replication and life, but also the phenomena of neuronal computing) and finally the processes of Darwinian evolution and genetics so brilliantly presented by Dawkins, most observers would agree that we've no doubt covered many of the basic essentials of a scientific outlook on reality. Deutsch himself refers to these four grand dimensions of his global philosophical approach as strands (a word I like, which evokes weaving a fabric).

A few weeks ago, I was excited to learn that Deutsch has been working on a kind of fifth strand, of a subterranean nature, which he calls constructor theory. If you've got 47 minutes of free time, I urge you to click here to listen to Deutsch himself presenting this work. Basically, it's a matter of trying to understand why certain things are possible (even though they may have never actually happened yet) whereas countless other potential events are impossible because certain laws of physics have "blacklisted" them forever. In other words, he has enhanced astronomically the sense of the concept of possibility, to the point of claiming that anything and everything is strictly possible... provided only that we know of no law of physics that forbids such a happening, and therefore renders it impossible. Deutsch draw our attention to the strict binarity of the situation. Between the impossible (ruled out by physics) and the possible, there is no third way out. On the one hand, nothing—not even the most extravagant events—should be branded as theoretically impossible unless we are already aware of a law of physics that forbids such things. On the other hand, everything else should be thought of as theoretically possible.

In his eagerness to point out the counterintuitive nature of this thinking, Deutsch hit upon an amusing easy-to-grasp example, which goes straight to the heart of my Antipodes blog. Most of us agree that people on the other side of the planet Earth are in an upside-down position with respect to us, and vice versa.

That old Epinal image is funny but quite silly, of course, because nobody really believes that Antipodeans get around on their hands, with their Hobbit-like feet stretching towards the heavens. But are we truly ready to admit that the heads of Antipodeans point constantly in the opposite direction to our own heads? If technology were to offer me a magical real-time closeup view of Antipodeans, in strict conformity with our mutual orientations, in the same way that binoculars enable me to observe distant objects through my bedroom window, would I not be somewhat surprised to receive upskirt images of Antipodean ladies whose heads appear to be receding upwards into the sky? My surprise (which would be inevitable, I think) would seem to confirm that, to a certain extent, I've never really believed wholeheartedly that the heads of Antipodeans point in the opposite direction to mine. And David Deutsch considers that this mild form of surprise, or doubt, reflects my persistent quest for a third way out, between the possible and the impossible. My scientific culture persuades me that there is no law of physics that would forbid Antipodeans from getting around in an upside-down position with respect to me. So, I conclude that it's perfectly possible for this to be the case. At the same time, I consider that modern laws of celestial mechanics have quashed forever all remnants of flat-Earth theories, meaning that it's unthinkable that the heads of Antipodeans might point in the same direction as mine. And yet I don't seem to have gone one tiny step further and admitted explicitly, in a tangible concrete sense, that people down on the opposite side of the planet are truly presenting me constantly (if only the Earth were transparent) with an upskirt vision of their environment.

What David Deutsch seems to be saying (in a roundabout fashion) is that we would do well to consider, in an equally tangible and concrete sense, that we exist within a multiverse where the quantum effects admitted by today's laws of physics must be thought of, not only as possible happenings, but as garden-variety aspects of the fabric of everyday reality. And I'm not sure that many of us are prepared, at present, to assimilate profoundly that weird mode of looking at existence. Between archaic fairy tales (often supported by so-called commonsense) and hard state-of-the-art science, we persist in hoping, if not believing, that there must surely be some kind of convenient "third way out".

Saturday, June 23, 2012

Elusive Turing

Today is the centenary of the birth of Alan Turing [1912-1954]. He is represented in Manchester by this park-bench sculpture, which includes the cyanide-laced apple that killed the genius.

Google has celebrated the centenary by creating an ingenious doodle representing a Turing machine, but it takes some time and effort to figure out what it's supposed to do.

In my earlier blog post entitled Turing that unknown [display], I suggested that it's not easy to grasp what exactly Turing achieved. Fortunately, the US computer-science author Charles Petzold has offered us an excellent book, The Annotated Turing, which explains precisely the achievements of Turing.

While it's true that Turing's contribution to the British war effort at Bletchley Park was invaluable, his achievements in code-breaking were not the reason why we consider Turing today as the patriarch of computing. Likewise, while we appreciate Turing's suggestion about considering convincing man/machine conversations as a criterion for so-called artificial intelligence, this too was not really an all-important factor in Turing's claim to fame. So, why is Turing so greatly admired by computer scientists?

Well, his invention of the abstract concept of a so-called Turing machine (like the one in the Google doodle) threw light upon the limitations of algorithmic devices such as computers. More precisely, to use a horrible German term, Turing demonstrated that the Entscheidungsproblem cannot be solved. And what is this exotic beast? You might call it the "mission accomplished" problem. Like George W Bush with his war games, computers will remain forever incapable of determining beforehand whether or not a certain computing challenge can indeed be handled successfully. Turing taught us that the only way of knowing whether or not a computer can handle such-and-such a complex challenge is to set the machine into action and see whether or not it soon halts with a solution.

You might say that Turing proved that the proof of the computer pudding is in the computing.

Monday, May 21, 2012

Turing, that unknown

In a month's time, the computing world will be celebrating the centenary of Alan Turing, who was born in London on 23 June 1912.

Much mystery still surrounds the life and work of this great Englishman, who can truly be considered as the founder of computer science. He designed a marvelous computing device that soon became known as the Turing Machine. It's so powerful and precise that it can perform any calculation whatsoever, no matter how complex, including those that are carried out today by the giant supercomputers used in space engineering, military calculations or meteorological predictions. My book Machina Sapiens, published in 1976, offered French readers (no doubt for the first time ever) a drawing of a Turing Machine, accompanied by demonstrations of how it worked.

But I don't have the impression that anybody went out and actually built such a machine... to handle his office accounting, say. In fact, although a simple Turing Machine can indeed perform any of the computations executed by a modern computer, I have to be truthful and point out that I wouldn't advise anybody to get involved in trying to use a Turing Machine to build a spreadsheet, say, or to carry out some word processing. And I'm even less certain that a Turing Machine would be an efficient tool for tweeting, or sending e-mails, or linking up to Facebook. The problem, you see, is that Turing Machines have to be programmed from scratch, and even the simplest tasks—such as the multiplication of two numbers—would represent a huge programming challenge. What's more, I'm not sure that anybody has ever bothered to actually build an operational Turing Machine. Like the model that illustrated my book, Turing Machines tend to remain on paper, on the pages of textbooks for computer science students.

So, why all the fuss about Alan Turing having invented a machine that is the grand-daddy of all computers, past, present and future? Well, it's a bit like Einstein's E = mC2. This simple equation was the key to understanding that matter can be transformed into energy. But, between understanding the equation and being able to obtain energy from a nuclear reactor, a lot of hard work needs to be carried out. You might say poetically that the Turing Machine defines the "soul" of any imaginable computer in the real world. But, to move from the abstract "soul" to a real "flesh and blood" computer, you have to envisage a huge amount of design, engineering and programming... of both a hardware and a software kind.

Funnily enough, although the Turing Machine can indeed carry out any imaginable task that might be performed by modern computers, it's greatest interest was that it enabled Turing and other logicians to discover that certain tasks could never be carried out by any imaginable computer whatsoever. For example, it is impossible for a computer to determine beforehand, when faced with certain algorithms, that it will indeed be able to reach the intended end of the algorithm and provide an answer. In this way, the Turing Machine appeared on the scene as a mechanical variant of the themes of incompleteness and undecidability elucidated mathematically by Kurt Gödel (seen in the following photo alongside Albert Einstein):

Gödel was still alive in the early '70s when I was visiting the USA in order to organize my future series of TV programs on the subject of men and machines (basically, artificial intelligence and brain research). I spoke on the phone with Gödel for 20 minutes or so, trying vainly to get him to agree to being interviewed for my TV project. But he insisted—no doubt sincerely—that he himself did not consider that his theorems had any significance whatsoever in modern society... that's to say, at the level of ordinary folk who watch TV. Maybe he was right.

Getting back to Turing, his most concrete claim to fame was surely the wizard-level code-breaking operations that he performed for the British government during World War II, at Bletchley Park.

He was a practicing homosexual at a time in the UK when relationships of this kind were branded as criminal. The poor man, suffering no doubt from a form of autism (Asperger Syndrome) that made him socially awkward, was obliged to undergo ignominious chemical castration. In June 1954, a fortnight before his 42nd birthday, Turing was found dead in his laboratory, poisoned by cyanide, and clutching a half-eaten apple.

A British journalist once asked Steve Jobs if the logo of Apple computer was intended as a tribute to Turing. "No, that's not the case," replied Jobs, "but God, we wish it were."

Thursday, July 26, 2007

Brilliant book

This excellent book by the Oxford physicist David Deutsch came out a decade ago, but I've only just got around to reading it. Seeking to lay the foundations of a vast theory of everything, Deutsch introduces four great domains of knowledge that he refers to as strands:

— Quantum physics

— Epistemology, inspired by the work of Karl Popper

— Theory of computation, inspired by the work of Alan Turing

— Theory of evolution, inspired by the work of Richard Dawkins.

It's rare to find an eclectic author who's prepared to blend such different disciplines into a synthetic whole.