Skip to main content

Artificial Intelligence - Melanie Mitchell *****

As Melanie Mitchell makes plain, humans have limitations in their visual abilities, typified by optical illusions, but artificial intelligence (AI) struggles at a much deeper level with recognising what's going on in images. Similarly in some ways, the visual appearance of this book misleads. It's worryingly fat and bears the ascetic light blue cover of the Pelican series, which since my childhood have been markers of books that were worthy but have rarely been readable. This, however, is an excellent book, giving a clear picture of how many AI systems go about their business and the huge problems designers of such systems face.

Not only does Mitchell explain the main approaches clearly, her account is readable and engaging. I read a lot of popular science books, and it's rare that I keep wanting to go back to one when I'm not scheduled to be reading it - this is one of those rare examples.

We discover how AI researchers have achieved the apparently remarkable abilities of, for example, the Go champion AlphaGo, or the Jeopardy! playing Watson. In each case these systems are tightly designed for a particular purpose and arguably have no intelligence in the broad sense.  As for what's probably the most impressively broad AI application of modern times, self-driving cars, Mitchell emphasises how limited they truly are. Like so many AI applications, the hype far exceeds the reality - when companies and individuals talk of self-driving cars being commonplace in a few years' time, it's quite clear that this could only be the case in a tightly controlled environment.

One example, that Mitchell explores in considerable detail are so-called adversarial attacks, a particularly AI form of hacking where, for example, those in the know can make changes to images that are invisible to the human eye but that force an AI system to interpret what they are seeing as something totally different. It's a sobering thought that, for example, by simply applying a small sticker to a stop sign on the road - unnoticeable to a human driver - an adversarial attacker can turn the sign into a speed limit sign as far as an AI system is concerned, with potentially fatal consequences.

Don't get me wrong, Mitchell, a professor of computer science who has specialised in AI, is no AI luddite. But unlike many of the enthusiasts in the field (or, for that matter, those who are terrified AI is about to take over the world), she is able to give us a realistic, balanced view, showing us just how far AI has to go to come close to the more general abilities humans make use of all the time even in simple tasks. AI does a great job, for example, in something like Siri or Google Translate or unlocking a phone with a face - but AI systems still have no concept of, for example, understanding (as opposed to recognising) what is in an image. Mitchell makes it clear that where systems learn from large amounts of data, it is usually impossible to uncover how they are making decisions (which makes the EU's law requiring transparent AI decisions pretty much impossible to implement), so we really shouldn't trust them with important outcomes as they could easily be basing their outcomes on totally irrelevant inputs.

Apart from occasionally finding the explanations of the workings of types of neural network a little hard to follow, the only thing that made me raise an eyebrow was being told that Marvin Minsky 'coined the phrase "suitcase word"' - I would have thought 'derived* the phrase from Lewis Carroll's term "portmanteau word"' would have been closer to reality.

There have been good books on the basics of AI already, and excellent ones on the problems that 'deep learning' and big data systems throw up. But without a doubt, Mitchell's book sets a new standard in giving an understanding of what's possible and how difficult it is to go further. It should be read by every journalist, PR person and politician before they pump out yet more hype on the AI future. Recommended.

* Polite term
Hardback:   

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

Comments

Post a Comment

Popular posts from this blog

Roger Highfield - Stephen Hawking: genius at work interview

Roger Highfield OBE is the Science Director of the Science Museum Group. Roger has visiting professorships at the Department of Chemistry, UCL, and at the Dunn School, University of Oxford, is a Fellow of the Academy of Medical Sciences, and a member of the Medical Research Council and Longitude Committee. He has written or co-authored ten popular science books, including two bestsellers. His latest title is Stephen Hawking: genius at work . Why science? There are three answers to this question, depending on context: Apollo; Prime Minister Margaret Thatcher, along with the world’s worst nuclear accident at Chernobyl; and, finally, Nullius in verba . Growing up I enjoyed the sciencey side of TV programmes like Thunderbirds and The Avengers but became completely besotted when, in short trousers, I gazed up at the moon knowing that two astronauts had paid it a visit. As the Apollo programme unfolded, I became utterly obsessed. Today, more than half a century later, the moon landings are
Post a Comment
Read more

Space Oddities - Harry Cliff *****

In this delightfully readable book, Harry Cliff takes us into the anomalies that are starting to make areas of physics seems to be nearing a paradigm shift, just as occurred in the past with relativity and quantum theory. We start with, we are introduced to some past anomalies linked to changes in viewpoint, such as the precession of Mercury (explained by general relativity, though originally blamed on an undiscovered planet near the Sun), and then move on to a few examples of apparent discoveries being wrong: the BICEP2 evidence for inflation (where the result was caused by dust, not the polarisation being studied),  the disappearance of an interesting blip in LHC results, and an apparent mistake in the manipulation of numbers that resulted in alleged discovery of dark matter particles. These are used to explain how statistics plays a part, and the significance of sigmas . We go on to explore a range of anomalies in particle physics and cosmology that may indicate either a breakdown i
Post a Comment
Read more

Splinters of Infinity - Mark Wolverton ****

Many of us who read popular science regularly will be aware of the 'great debate' between American astronomers Harlow Shapley and Heber Curtis in 1920 over whether the universe was a single galaxy or many. Less familiar is the clash in the 1930s between American Nobel Prize winners Robert Millikan and Arthur Compton over the nature of cosmic rays. This not a book about the nature of cosmic rays as we now understand them, but rather explores this confrontation between heavyweight scientists. Millikan was the first in the fray, and often wrongly named in the press as discoverer of cosmic rays. He believed that this high energy radiation from above was made up of photons that ionised atoms in the atmosphere. One of the reasons he was determined that they should be photons was that this fitted with his thesis that the universe was in a constant state of creation: these photons, he thought, were produced in the birth of new atoms. This view seems to have been primarily driven by re
Post a Comment
Read more