Deep Learning: John Kelleher **

This is an entry in a series from the MIT Press that selects a small part of a topic (in this case, a subset of artificial intelligence) and gives it an 'essential knowledge' introduction. The problem is, there seems to be no consistency over the target audience of the series.

I previously reviewed Virtual Reality in the same series and it kept things relatively simple and approachable to the general reader, even if it did overdo the hype. This book by John Kelleher starts gently, but by about half way through it has become a full-blown simplified textbook with far too much in-depth technical content. That's exactly what you don't want in a popular science title.

What we get is plenty of detail of what deep learning-based systems are and how they work at the technical level, but there is practically nothing on how they fit with applications (unless you count playing games), which are described but not really explained, nor is there anything much on the problems that arise when deep learning is used for real world applications. There is a passing reference, admittedly to the difficulties of understanding how a deep learning AI system came to a decision and how this clashes with the EU's GDPR requirement for transparency and explanation, but if feels more like this is done to criticise the naivety of the legislation than the danger of using such systems.

Similarly, I saw nothing about the dangers of deep learning systems using big data picking up on correlations that don't involve any causal link, nor does the book discuss the long tail problems that arise with inputs that are relatively uncommon and so are unlikely to turn up in the training data. Similarly we read nothing about the dangers of adversarial attacks, which can fool the systems into misinterpreting inputs with tiny changes, or the difficulties such systems have with real, messy environments as opposed to the rigid rules of a game.

Overall, the book is both pitched wrong and doesn't cover the aspects that really matter to the public. It may well do fine as an introductory text for a computer science student, but that doesn't fit with the blurb on the back, which implies it is for public consumption.

Paperback:

Kindle:

Using these links earns us commission at no cost to you

Review by Brian Clegg

Comments

Luna: Moon Rising (SF) - Ian McDonald ****

I'm not the natural audience for this book. Game of Thrones l eaves me cold - and it's hard not to feel the influence of GoT (and a whole lot of Dune ) underneath a veneer of science fiction and the trappings of a South American drug cartel in the cod-medieval family power battles and chivalric details. There are even dragons (of a sort). I'd be really sad if the future did involve this sort of throwback feudalism. However, remarkably, despite this I found Luna: Moon Rising kept me engaged. The fact is that Ian McDonald can put together a good plot with intricate machinations, which is enough to carry the reader through what can be a bewildering collection of characters. The two page scene-setter saying who did what to whom at the start was useful, but I could have done with family trees for the main family as I was constantly forgetting who was who - especially easy as McDonald endows many families with characters with the same first initial (e.g. Ariel and Al...

Adventures of a Computational Explorer - Stephen Wolfram ***

Stephen Wolfram, the man behind the scientist's mathematical tool of choice, Mathematica, plus a whole host of other software products, including the uncanny Wolfram Alpha knowledge engine, is undoubtedly a genius of the first order. In this book, we get an uncensored excursion into the mind of genius - which is, without doubt, a fascinating prospect. The book consists of a collection of essays and speeches that Wolfram has produced over the last ten to fifteen years, covering an eclectic range of topics. Like all such collections, the result is something that lacks the coherence of a book with a narrative that runs through it, inevitably introducing a degree of repetition and a mix of interesting and not-so-interesting topics - but there's likely to be something to catch the attention anyone who is into computing or mathematics. One of the most interesting pieces is the opening one, where Wolfram describes being a consultant on the SF movie Arrival. He seems to hav...

E=mc2: A biography of the world’s most famous equation – David Bodanis *****

David Bodanis is a storyteller, and he fulfils this role with flair in E=mc2. The premise of the book is simple – Einstein himself has been biographed (biographised?) to death, but no one has picked out this most famous of equations, dusted it down and told us what it means, where it comes from and what it has delivered. Allegedly, Bodanis was inspired to write the book after hearing see an interview with actress Cameron Diaz in which she commented that she’d really like to know what that famous collection of letters was all about. Although the book had been around for a while already when this review was written (September 2005), it seemed a very apt moment to cover it, as the equation is, as I write, exactly 100 years old. So when better to have a biography? Bodanis starts off by telling us about the individual elements of the equation. What the different letters mean, where the equal sign comes from and so on. This is entertaining, though he seems to tire of the approach on...

Search This Blog