Skip to main content

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

Popular posts from this blog

Six Maths Essentials - Terence Tao *****

This is a book in the 'compact collection of science/maths essays' category beloved of publishers because they can charge hardback prices for fewer than 80 pages of text - but don't let that put you off, because it's excellent. Terence Tao introduces six areas of mathematics: numbers, algebra, geometry, probability, analysis and dynamics. Each of these is descriptive rather than practical (there are only a couple of school-level equations in sight), done in a way that gets to the heart of 'what's the point of this subject' with precision and a good level of storytelling. It's a shame in a way that analysis and dynamics form the final third of the book (apart from a short epilogue), because they don't work as well as the other sections. For the first four, practically everyone will have a vague idea of what they are and what they do - but Tao is then able to open our eyes to their wonders and power. Because Tao has to stay away from any detail of what...

Bradley Efron - Five Way Interview

Bradley Efron is Max Stein Professor emeritus of Statistics and of Biomedical Data Science at Stanford University. He is the inventor of the bootstrap method for assessing statistical accuracy and has published extensively on the theory and practice of data analysis. He has served as president of the American Statistical Association and the Institute of Mathematical Statistics.  A MacArthur fellow, he is a member of the National Academy of Sciences and received the National Medal of Science in 2005. His remarkable contributions to his field were recognized again with the 2018 International Prize in Statistics, which noted his influence on 'statistical thinking in every way, including new pathways in how we communicate and teach statistics to all.' His latest book is To Think Like a Statistician . Why mathematics?   Abstract mathematics and I had an amicable divorce following my Caltech graduation. (I really wasn’t very good at it.) Statistics is a branch of applied mathematics...

To Think Like a Statistician - Bradley Efron ****

An old advert had the tag line 'it does what it says on the tin'. There's a vagueness about the title of this book that requires Bayes' theorem to work out whether it really does what it says on the cover. If you think this book will tell you how to think like a statistician, it's unlikely that you will be satisfied. But if you think it will tell you how a statistician thinks, it is much closer to being true. Sadly this makes the tag line I've seen 'Why you don’t need to be a statistician to think like one' entirely wrong. In the book, Bradley Efron gives us plenty of detail on how statisticians apply probability to try to understand better what is happening in the world. He makes a great point that his high school mathematical education consisted primarily of algebra, geometry and trigonometry. While these 'proved invaluable to my subsequent scientific studies' they hardly ever crop up in the news, where the one mathematical subject that does oc...