Skip to main content

How Smart Machines Think - Sean Gerrish ****

While it will become apparent I think this book should have been titled 'How Dumb Machines Think', it was a remarkably enjoyable insight into how the well publicised AI successes - self-driving cars, image and face recognition, IBM's Jeopardy! playing Watson, along with  game playing AIs in chess, Go and Atari and StarCraft, perform their dark arts.

There's no actual programming presented here, so no need for non-programmers to panic, though there is some quite detailed discussion of how the software architectures are structured and how the different components - for example neural networks - do their job, but it isn't anything too scary if you take it slowly.

One thing that comes across very strongly, despite the AI types' insistence that their programs are of general use, is how very specifically tailored programs like the AlphaGo software that beat champions at the game Go, and the Watson computer that won at the US TV quiz show Jeopardy! were - incredibly finely designed to meet use and that use only.

The reason I make the remark about dumb machines is that what doesn't come across sufficiently in Sean Gerrish's book is that, because these programs are not in any sense intelligent, when they get things wrong, they often get things dramatically wrong. So some of Watson's answers on Jeopardy! did not make any sense at all. Similarly, image recognition software can be fooled by apparently abstract patterns that happen to have the right components to appear to be a distinguishable object. And when you bear in mind we're suggesting putting this kind of software in charge of cars that 'getting it dramatically wrong' bit is more than a little unnerving.

There was, though, a great section on the development of self-driving cars, from the original feeble attempts, where all the competitors in a race failed before completing 10 percent of the course, through to more recent and more successful versions that can handle basic traffic scenarios - though it would have been nice if Gerrish had gone beyond the old prize challenges to describe what the latest Google and Uber vehicles do. (It may be that their approaches are too proprietary.)

However, what was missing was any serious assessment of the big problems still faced. There have been two excellent books recently on the huge holes in AI that practitioners rarely admit to - The AI Delusion and Common Sense, The Turing Test and the Quest for Real AI - Gerrish would have produced an even better book if he could have addressed the concerns that these books raise.

Even so, in How Smart Machines Think we have a hugely informative and very readable book for anyone with an interest in finding out just what the much-trumpeted AI systems really do, and what lies beneath the hype.

Hardback:  

Kindle:  
Using these links earns us commission at no cost to you

Review by Brian Clegg

Comments

Popular posts from this blog

On the Fringe - Michael Gordin *****

This little book is a pleasant surprise. That word 'little', by the way, is not intended as an insult, but a compliment. Kudos to OUP for realising that a book doesn't have to be three inches thick to be interesting. It's just 101 pages before you get to the notes - and that's plenty. The topic is fringe science or pseudoscience: it could be heavy going in a condensed form, but in fact Michael Gordin keeps the tone light and readable. In some ways, the most interesting bit is when Gordin plunges into just what pseudoscience actually is. As he points out, there are elements of subjectivity to this. For example, some would say that string theory is pseudoscience, even though many real scientists have dedicated their careers to it. Gordin also points out that, outside of denial (more on this a moment), many supporters of what most of us label pseudoscience do use the scientific method and see themselves as doing actual science. Gordin breaks pseudoscience down into a n

A (Very) Short History of Life on Earth - Henry Gee *****

In writing this book, Henry Gee had a lot to live up to. His earlier title  The Accidental Species was a superbly readable and fascinating description of the evolutionary process leading to Homo sapiens . It seemed hard to beat - but he has succeeded with what is inevitably going to be described as a tour-de-force. As is promised on the cover, we are taken through nearly 4.6 billion years of life on Earth (actually rather more, as I'll cover below). It's a mark of Gee's skill that what could have ended up feeling like an interminable list of different organisms comes across instead as something of a pager turner. This is helped by the structuring - within those promised twelve chapters everything is divided up into handy bite-sized chunks. And although there certainly are very many species mentioned as we pass through the years, rather than feeling overwhelming, Gee's friendly prose and careful timing made the approach come across as natural and organic.  There was a w

Michael D. Gordin - Four Way Interview

Michael D. Gordin is a historian of modern science and a professor at Princeton University, with particular interests in the physical sciences and in science in Russia and the Soviet Union. He is the author of six books, ranging from the periodic table to early nuclear weapons to the history of scientific languages. His most recent book is On the Fringe: Where Science Meets Pseudoscience (Oxford University Press). Why history of science? The history of science grabbed me long before I knew that there were actual historians of science out there. I entered college committed to becoming a physicist, drawn in by the deep intellectual puzzles of entropy, quantum theory, and relativity. When I started taking courses, I came to understand that what really interested me about those puzzles were not so much their solutions — still replete with paradoxes — but rather the rich debates and even the dead-ends that scientists had taken to trying to resolve them. At first, I thought this fell under