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
Labels:

Comments

Post a Comment

Popular posts from this blog

Beyond Belief - Helen Pearson *****

Apparently it comes as a surprise to many that medicine was not particularly scientific until the end of the twentieth century (to be honest, it's no surprise to me - we had a GP who used homeopathy in the 90s). Instead it was based on anecdotal guidance - the kind of thing that appeared to work. Evidence-based medicine has since improved the field, trying where possible to base decisions on evidence, ideally based on randomised controlled trials. The first part of Helen Pearson's book covers this well - though I think it's by far the least interesting part of what we discover. Instead what's truly fascinating is the rest of it, looking at a wide range of other fields where evidence was rarely properly used and that are only now starting to dip a toe in the water. These include social policy, policing, conservation, business and education. The main part of the book gives us examples of how bad these areas have been in terms of basing decisions on what's always been ...
Post a Comment
Read more

In Seach of Sea Dragons - Matthew Myerscough ****

It's common advice to would-be authors of narrative non-fiction to open with something dramatic - Matthew Myerscough certainly does this with the story of his being trapped under an avalanche on Snowdon (while his girlfriend, also carried away remains on top of the snow unhurt). It certainly is dramatic, but seemed entirely disconnected from the reason I got the book, which was to read about fossil collecting.  Luckily, though, in the second chapter we get into a more conventional 'how I got interested in fossils as a boy'. Having recently reviewed Patrick Moore's autobiography and noting that astronomy was one of the few sciences where amateurs can still make a contribution, it came to mind that palaeontology is another - Myerscough is a civil engineer by trade, but just as amateur astronomers can find new details in the skies, so amateur fossil hunters have been searching for these relics for centuries. When I give talks in junior schools, the two topics that guarant...
Post a Comment
Read more

The Infinity Machine - Sebastian Mallaby ****

It's very quickly clear that Sebastian Mallaby is a huge Demis Hassabis fan - writing about the only child prodigy and teen genius ever who was also a nice, rounded personality. After a few chapters, though, things settle down (I'm reminded of Douglas Adams' description of the Hitchhiker's Guide to the Galaxy ) and we get a good, solid trip through the journey that gave us DeepMind, their AlphaGo and AlphaFold programs, the sudden explosion of competition on the AI front and thoughts on artificial general intelligence. Although Mallaby does occasionally still go into fan mode - reading this you would think that AlphaFold had successfully perfectly predicted the structure of every protein, where it is usually not sufficiently accurate for its results to have direct practical application - we get a real feel for the way this relatively unusual company was swiftly and successfully developed away from Silicon Valley. It's readable and gives an important understanding of...
Post a Comment
Read more