Skip to main content

Common Sense, The Turing Test and the Quest for Real AI - Hector Levesque *****

It was fascinating to read this book immediately after Ed Finn's What Algorithms Want. They are both by academics on aspects of artificial intelligence (AI) - but where reading Finn's book is like wading through intellectual treacle, this is a delight. It is short, to the point, beautifully clear and provides just as much in the way of insights without any of the mental anguish.

The topic here is the nature of artificial intelligence, why the current dominant approach of adaptive machine learning can never deliver true AI and what the potential consequences are of thinking that learning from big data is sufficient to truly act in a smart fashion.

As Hector Levesque points out, machine learning is great at handling everyday non-exceptional circumstances - but falls down horribly when having to deal with the 'long tail', where there won't be much past data to learn from. For example (my examples, not his), a self-driving car might cope wonderfully with typical traffic and roads, but get into a serious mess if a deer tries to cross the motorway in front of it, or should the car encounter Swindon's Magic Roundabout.

There is so much here to love. Although the book is compact (and rather expensive for its size), each chapter delivers excellent considerations. Apart from the different kinds of AI (I love that knowledge-based AI has the acronym of GOFAI for 'good old-fashioned AI'), this takes us into considerations of how the brain works, the difference between real and fake intelligence, learning and experience, symbols and symbol processing and far more. Just to give one small example of something that intrigued me, Levesque gives the example of a very simple computer program that generates quite a complex outcome. He then envisages taking the kind of approaches we use to try to understand human intelligence - both psychological and physiological - showing how doing the same thing with this far simpler computer equivalent would fail to uncover what was happening behind the outputs.

For too long, those of us who take an interest in AI have been told that the 'old-fashioned' knowledge-based approach was a dead end, while the modern adaptive machine learning approach, which is the way that, for instance, programs like Siri and Alexa appear to understand English, is the way forward. But as the self-driving car example showed above, anything providing true AI has to be reliable and predictable to be able to cope with odd and relatively unlikely circumstances - because while any individual unlikely occurrence will probably never happen, the chances are that something unlikely will come along. And when it does, it takes knowledge to select the most appropriate action.

Highly recommended.

Hardback:  

Kindle 
Review by Brian Clegg

Comments

Popular posts from this blog

The Demon in the Machine - Paul Davies *****

Physicists have a habit of dabbling in biology and, perhaps surprisingly, biologists tend to be quite tolerant of it. (I find it hard to believe the reverse would be true if biologists tried to do physics.) Perhaps one reason for that tolerance is Schrödinger’s lecture series and book What is Life?, which had a huge impact on molecular biology and with a reference to which, not surprisingly, Paul Davies begins his fascinating book. 

At the heart of the The Demon in the Machine (we'll come back to that demon in a moment) is the relationship between life and information. In essence, Davies points out that if we try to reduce life to its simple physical components it is like trying to work with a computer that has no software. The equivalent of software here is information, not just in the best publicised aspect of the information stored in the DNA, but on a far broader scale, operating in networks across the organism.
This information and its processing gives life its emergent compl…

The Cosmic Mystery Tour – Nicholas Mee ****

This is another book, like last year’s Enjoy Our Universe by Alvaro de Rújula, that sets out to provide a light-hearted introduction to physics and astrophysics for the general reader. It’s from the same publisher (OUP) and packaged in the same way: as a high quality small-format hardback with 200 glossy pages, the majority of them adorned with colour pictures. But that’s where the resemblance ends. Unlike its predecessor, this new book by Nicholas Mee delivers exactly what it promises.

It’s not that de Rújula’s book was a bad one, but he just wasn’t able to think his way into the reader’s mind. He kept saying ‘physics is fun’, but he was talking about the fun a professional physicist gets out of doing it – which is a very arcane, often highly mathematical, type of fun. The result, for a non-specialist reader, was actually quite alienating. Mee, on the other hand, understands exactly how his readers think, what they find interesting, and the details that – no matter how important they …

Professor Maxwell's Duplicitous Demon - Brian Clegg ****

‘It’s not uncommon when trying to give Maxwell his rightful place in the pantheon of physics to bracket him with Newton and Einstein’, Brian Clegg says towards the end of this book. In one sense that’s perfectly true. Dip into any physics textbook and you’ll see Maxwell’s name at least as often as the other two. His greatest achievement – Maxwell’s equations – did for electromagnetism what Newton had done for gravity, while laying the essential theoretical groundwork for everything Einstein was to do.

There’s a big difference, though. A few years ago, when I was offered the chance to write short biographies of Newton and Einstein, I jumped at it – because they addressed mysteries of the universe that anyone can relate to, and their lives outside physics were, if anything, even more fascinating. At the risk of sounding downright rude, you can’t say either of those things about James Clerk Maxwell. In spite of that, Brian Clegg has done a wonderful job here of recounting just what Maxwel…