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

The God Game (SF) - Danny Tobey *****

Wow. I'm not sure I've ever read a book that was quite such an adrenaline rush - certainly it has been a long time since I've read a science fiction title which has kept me wanting to get back to it and read more so fiercely. 

In some ways, what we have here is a cyber-SF equivalent of Stephen King's It. A bunch of misfit American high school students face a remarkably powerful evil adversary - though in this case, at the beginning, their foe appears to be able to transform their worlds for the better.

Rather than a supernatural evil, the students take on a rogue AI computer game that thinks it is a god - and has the powers to back its belief. Playing the game is a mix of a virtual reality adventure like Pokemon Go and a real world treasure hunt. Players can get rewards for carrying out tasks - delivering a parcel, for example, which can be used to buy favours, abilities in the game and real objects. But once you are in the game, it doesn't want to let you go and is …

Uncertainty - Kostas Kampourakis and Kevin McCain ***

This is intended as a follow-on to Stuart Firestein's two books, the excellent Ignorance and its sequel, Failure, which cut through some of the myths about the nature of science and how it's not so much about facts as about what we don't know and how we search for explanations. The authors of Uncertainty do pretty much what they set out to do in explaining the significance of uncertainty and why it can make it difficult to present scientific findings to the public, who expect black-and-white facts, not grey probabilities, which can seem to some like dithering.

However, I didn't get on awfully well with the book. A minor issue was the size - it was just too physically small to hold comfortably, which was irritating. More significantly, it felt like a magazine article that was inflated to make a book. There really was only one essential point made over and over again, with a handful of repeated examples. I want something more from a book - more context and depth - that …

The Ascent of Gravity - Marcus Chown ****

Marcus Chown is one of the UK's best writers on physics and astronomy - it's excellent to see him back on what he does best. Here we discover our gradual approach to understanding the nature of gravity - the 'ascent' of the title - which, though perhaps slightly overblown in the words 'the force that explains everything' (quantum physics does quite a lot too, for example), certainly makes us aware of the importance of this weakest of fundamental forces. Chown's approach to gravity is a game of three halves, as they say, broadly covering Newton, Einstein and where we go from general relativity.
As far as the first two sections go, with the exception of the 2015 gravitational waves detection, there's not much that's actually new - if you want a popular science exploration of these aspects of the topic with more depth see this reviewer's Gravity - but no one has covered the topic with such a light touch and joie de vivre as Chown. 
Although Chown doe…