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

De/Cipher - Mark Frary ****

I was a little doubtful when I first saw this book. Although it has the intriguing tagline 'The greatest codes ever invented and how to crack them' the combination of a small format hardback and gratuitous illustrations made me suspect it would be a lightweight, minimal content, Janet and John approach to codes and ciphers. Thankfully, in reality Mark Frary manages to pack a remarkable amount of content into De/Cipher's slim form.

Not only do we get some history on and instructions to use a whole range of ciphers, there are engaging little articles on historical codebreakers and useful guidance on techniques to break the simpler ciphers. The broadly historical structure takes the reader through basic alphabetic manipulation, keys, electronic cryptography, one time pads and so on, all the way up to modern public key encryption and a short section on quantum cryptography. 

We even get articles on some of the best known unsolved ciphers, such as the Dorabella and the Voynich ma…

Cosmology for the Curious - Delia Perlov and Alex Vilenkin ***

In the recently published The Little Book of Black Holes we saw what I thought was pretty much impossible - a good, next level, general audience science title, spanning the gap between a typical popular science book and an introductory textbook, but very much in the style of popular science. Cosmology for the Curious does something similar, but coming from the other direction. This is an introductory textbook, intended for first year physics students, with familiar textbook features like questions to answer at the end of each chapter. Yet by incorporating some history and context, plus taking a more relaxed style in the writing, it's certainly more approachable than a typical textbook.

The first main section, The Big Bang and the Observable Universe not only covers basic big bang cosmology but fills in the basics of special and general relativity, Hubble's law, dark matter, dark energy and more. We then move onto the more speculative (this is cosmology, after all) aspects, brin…

Paul McAuley - Four Way Interview

Paul McAuley won the Philip K. Dick Award for his first novel and has gone on to win the Arthur C. Clarke, British Fantasy, Sidewise and John W. Campbell Awards. He gave up his position as a research biologist to write full-time. He lives in London. His latest novel is Austral.


Why science fiction?

For one thing, I fell in love with science fiction at an early age, and haven’t yet fallen out of love with it (although I have flirted with other genres). For another, we’re living in an increasingly science-fictional present. Every day brings headlines that could have been ripped from a science-fiction story. Giant robot battle: Who knew a duel between chainsaw-armed mech suits could be so boring? for instance. Or, Roy Orbison hologram to embark on UK tour in 2018. And looming above all this, like Hokusai’s famous wave, are the ongoing changes caused by global warming and climate change, which is just one consequence of human activity having become the dominant force of change on the planet…