Thursday, 20 April 2017

Four Way Interview - Hector Levesque

Hector Levesque is Professor Emeritus in the Department of Computer Science at the University of Toronto. He worked in the area of knowledge representation and reasoning in artificial intelligence. He is the co-author of a graduate textbook and co-founder of a conference in this area. He received the Computers and Thought Award in 1985 near the start of his career, and the Research Excellence Award in 2013 near the end, both from IJCAI (the International Joint Conferences on Artificial Intelligence). His latest title is Common Sense, The Turing Test, and the Quest for Real AI.

Why computer science?

Computer science is not really the science of computers, but the science of computation, a certain kind of information processing, with only a marginal connection to electronics. (I prefer the term used in French and other languages, informatics, but it never really caught on in North America.) Information is somewhat like gravity: once you are made aware of it, you realize that it is everywhere. You certainly cannot have a Theory of Everything without a clear understanding of the role of information. 

Why this book?

AI is the part of computer science concerned with the use of information in the sort of intelligent behaviour exhibited by people. While there is an incredible amount of buzz (and money) surrounding AI technology these days, it is mostly concerned with what can be learned by training on massive amounts of data. My book makes the case that this is an overly narrow view of intelligence, that what people are able to do, and what early AI researchers first proposed to study, goes well beyond this.

What's next?

I have a technical monograph with Gerhard Lakemeyer published in 2000 by MIT Press on the logic of knowledge bases, that is, on the relationship between large-scale symbolic representations and abstract states of knowledge. We are working on a new edition that would incorporate some of what we have learned about knowledge and knowledge bases since then. 

What's exciting you at the moment?

For me, the most exciting work in AI these days, at least in the theoretical part of AI, concerns the general mathematical and computational integration of logical and probabilistic reasoning seen, for example, in the work of Vaishak Belle. It's pretty clear to all but diehards that both types of knowledge will be needed, but previous solutions have been somewhat ad hoc and required giving up something out of one or the other.

Finding Fibonacci - Keith Devlin ***

I was rather surprised by this book, but I shouldn't have been. It does exactly what it says on the tin. The subtitle is 'The quest to rediscover the forgotten mathematical genius who changed the world.' So I shouldn't really complain that the book is far more about the quest than about the mathematician Fibonacci and his work - but I was disappointed nonetheless.

In practice, I enjoyed the details of the search for Fibonacci pointers like street signs, as it's the kind of thing I've done myself as a science writer. But I think Keith Devlin suffered from a common problem with someone who gets to close to a subject. If, for example, you are a birdwatcher and have spent ages tracking down a lesser spotted grebe, you might assume that the rest of us are as interested as you are - but we really aren't. There was just too much detail on Devlin's attempts to track down early copies (there are no originals) of Fibonacci's work. And where there is a tiny little bit of drama, he blows it out of all proportion with overselling. Take this passage:
What I would learn from that visit was that the story of the Liber abbaci [Fibonacci's oddly titled book introducing the numbers] is a very human one, spanning many centuries, with an ending (assuming the translation into English is its ending) every bit as dramatic as any Hollywood scriptwriter could dream up.
It really isn't. It's engaging without doubt, but not exactly worthy of the dramatic lead up. It didn't help that the book opens with an introduction that sounds like it was written by Troy McClure from the Simpsons (the actor character who introduces himself along the lines of 'You may remember me from such educational films as...') - Devlin spends quite a while telling us he is quite famous as 'the math guy'. Okay. 

At its core, there's some good material here about the introduction of our current numerals from India via the Arabic world through the influence of Fibonacci's book, which resulted in a whole chain of 'how to' smaller books for practical use. I don't doubt Devlin's assertion that the introduction of these numbers was crucial to finance, trade and science, though I think he over-inflates the importance of Fibonacci as an individual. The number system would have arrived anyway - and though he certainly had a strong influence on its use, it's interesting that in some countries there was strong resistance, with records having to use numbers written out as words for several centuries.

In a way, the problem with this book is it's a bit like one of those 'Making of' TV shows you get when a blockbuster film comes out. There are a lot of tantalising mentions of things in Devlin's 'real' book on Fibonacci, where this title focusses very much on his adventures visiting libraries in Italy. There's also a lot of repetition - sometimes almost word for word between chapters - all in all, it would have made a really good magazine article, but I'm not sure it's a book I'd recommend unless you are particularly interested in the details of doing this kind of research.

Review by Brian Clegg

Monday, 17 April 2017

Images from a Warming Planet - Ashley Cooper ****

I'm not a big fan of coffee table books - and this is certainly big and heavy enough to make a coffee table from if you attached legs (I certainly wouldn't fancy reading it in bed) - but Ashley Cooper's 495 photographs covering the impact of climate change, fossil fuel energy generation and renewables are absolutely stunning - crisp, large scale, extremely colourful and putting across a message that is sometimes hard to convey in words. Whether Cooper is showing oil refineries and damaged landscapes or homes battered by storms and decaying after drought, these are dramatic images.

In his foreword, Jonathon Porritt says 'do not allow the power of the images come between you and the people' - emphasising that it's not enough to record what's happening, but rather we should this kind of thing as a call to arms. My only concern here is that there's an awful lot of 'isn't this terrible' scenes and some nice images of the small scale 'good stuff' - but not enough suggestive of how we get from here to there. For example, there's a section on renewables which has a number of visually striking wind farms, but not much in the way of the more important solar farms - and no mention of the huge need to develop better storage (which often feels a bit too hard-tech to be truly green) because most renewables are not consistent in their output.

From a scientific viewpoint, I would have liked to have seen more restraint in the assumptions of causality. We are presented with the impact of storms, for example, as if they are definitively the result of climate change. Scientists would always temper the assertion, because there have always been devastating storms. We can say with a high probability that storms are more frequent as a result of climate change, but we can't say that a particular storm was caused by it. There's also an amusing little scientific contradiction on a page that comments about Iceland getting 30% of its energy from geothermal and also carries a picture of the Sun saying 'Ultimately all the Earth's energy comes from the Sun' - it certainly mostly does, but there are exceptions, notably nuclear (which only gets a passing dismissal later on) and geothermal.

So there may be rather too much of an 'isn't this terrible' message accompanied by over-simplistic solutions - but this shouldn't get in the way of these stunning images and the job I hope they can do in persuading some of the undecided on climate change. If the pictures get the message that we need to do something across to their viewership (somehow, readership doesn't work) and they then look elsewhere for those solutions, then at least the images are doing their job. And despite Mr Porritt's concerns, I don't think we should underplay the power of the photography. It's a great book visually and I'm delighted to have had a chance to get a good look at it.

The best place to get the book is the author's website:

Review by Brian Clegg

Saturday, 15 April 2017

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.

Review by Brian Clegg

Friday, 14 April 2017

What Algorithms Want - Ed Finn ****

The science fiction author Neal Stephenson comments on the cover of this book that it is 'highly enjoyable'. I suspect this is because in the opening of the book, Ed Finn repeatedly refers to Stephenson's impressive novel Snow Crash. If Stephenson actually found reading What Algorithms Want to be fun, he needs to get out more. I would, instead, describe it is extremely hard work to read - but it is hard work that is rewarded with some impressive insights. What Algorithms Want is both clever and able to cut away the glamour (in the old sense of the word) of the internet and the cyber world to reveal what's really going on beneath - as long as you can cope with the way that the book is written.

A Finn points out, most of us rely on algorithms, from Google’s search to Facebook’s timeline, not consciously considering that these aren’t just tools to help us, but processes that have their own (or their makers) intentions embodied in them. What’s more, there's really important material here about the insidious way the cyber world is driving us towards processes where the effort is not not concerned with a final product so much as the continuation of the system. However, there is also a fair amount of pretentious content that can be reminiscent of Sokal’s famous hoax - all too often, the sources Finn quotes seem to be using words from the IT world without entirely understanding them.

In terms of readability, What Algorithms Want suffers from same problem as the output of many of the university students I help with writing - ask them what their essay means and what they say is much clearer that what they've written. Finn is overly fond, for instance, of ‘fungible’ and ‘imbricate’, which are not words I'd really like to see outside of specialist publications.

Some of the specifics don't quite ring true. Finn talks about a modern equivalent of Asimov's fictional psychohistory, without covering the way that chaos theory makes it clear that an algorithmic representation of such a complex system could never produce useful predictions (any more than we can ever forecast the weather more than a few days into the future). Sometimes, Finn seems to be complaining about something that isn't a match to reality. So, for instance, he says 'you listen to a streaming music station that almost gets it right, telling yourself that these songs, not quite the right ones, are perfect for this moment because a magic algorithm selected them.' He absolutely misses the point. You don't do it because you think a magic algorithm produces perfection. You do it because you haven't time to spend an hour assembling the perfect playlist for the moment. It's a convenient compromise - and a far better match than listening to a random playlist off the radio.

This reflects a tendency to read too much into an example. For example, despite admitting that Netflix gave the makers of the series House of Cards carte blanche on content, Finn still identifies algorithmic aspects to the script. This comes out particularly when he comments at length on Fincher's idea of having Underwood address the audience direct, apparently not aware that this was one of the standout features of the (very non-algorithmic) original BBC series. Elsewhere, when talking about Uber, Finn says 'The company's opacity about pricing and the percentage of revenue shared with drivers makes it even more Iike an arbitrary video game' - but what conventional business does share this kind of information with its customers? Do you know the markup on products in your corner store, or what a McDonald's franchisee earns?

However, don't let these detailed complaints put you off. There is a huge amount to appreciate here, especially when Finn gets onto individual aspects of the impact of algorithms on our lives in the likes of Siri, Google and Netflix. And throughout there is much to challenge the reader, encouraging thought about technology we tend to take for granted. I just wish that it could have been written in a less obscure fashion.

Review by Brian Clegg

Monday, 10 April 2017

The Reality Frame - Brian Clegg *****

I’ve read quite a few of Brian Clegg’s books, but this one’s outstanding. Relativity is a topic that many writers struggle to get across - Clegg does this brilliantly thanks to two tactics I’ve never seen before. First, he makes use of that most fundamental requirement for relativity, the frame of reference. It’s not just the title of the book that suggests frames of reference - this concept forms the backbone of his exploration of relativity. But then he goes totally mad and builds a universe from scratch!

This audacious approach enables us to see, piece by piece, that relativity is about far more than Einstein’s work - fascinating though that is. It’s not that he ignores special and general relativity. There’s even an appendix where he shows how it only takes a maths GCSE to follow the mathematics that make time dilation happen. (I'd like to see more of this kind of thing in popular science books.) But he goes far beyond Einstein's work. For example, in the final chapters he introduces life and creativity to his universe and shows the essential roles that relativity and frames of reference have to play in those cases. 

In bringing in creativity, Clegg gives the book a human focus, and this then builds to a chance to reassess humanity’s place in the universe. The book mentions Bronowski's classic The Ascent of Man, which is a brave parallel to draw, but there are some real parallels in a very different kind of book. I thought I knew the basics of relativity - yet despite never becoming over-technical, The Reality Frame really opened my eyes to a different way of looking at the universe. Clegg quotes my favourite physicist, Richard Feynman on the laws of nature - this is a chance to see those laws in a new light.

Review by Peet Morris
Please note, this title is written by the editor of the Popular Science website. Our review is still an honest opinion – and we could hardly omit the book – but do want to make the connection clear.

Sunday, 9 April 2017

4th Rock from the Sun - Nicky Jenner ***

It's rather appropriate that the title of this book seems to be based on the name of the so-so TV show '3rd Rock from the Sun' - like that programme, it shows promise and has plenty of good bits, but as a whole it's rather confused, lacks structure and could have been so much better.

What we have here is a rag bag of everything the general reader might want to know about Mars - or about anything vaguely related to Mars. Nicky Jenner has an engagingly open writing style, although I think the copy editor at Bloomsbury must have been on holiday - I've never known a copy editor let a single exclamation mark through in an adult title, yet this one is littered with them! That did make it feel occasionally as if I was reading a book for children! But it's not!

Getting off my punctuation hobby horse, the good news is that there is lots of genuinely interesting material about Mars here, including stuff I've never seen in a popular science title before. Some of this is in the detail of the geology - how and why the two hemispheres are so different, and the future of the moons, for instance - and some is in the cultural impact of Mars. I've never read a book on the subject that doesn't bring in War of the Worlds as a cultural marker - but I can't remember any waxing lyrical on the subject of Captain Scarlet, a series from the makers of Thunderbirds which I loved as a child (though I had forgotten the Martian aspect until Jenner reminded me).

There's plenty more to like. Jenner takes a realistic view of the various potential attempts to get humans to Mars and points out very firmly how far the concept of terraforming strays into science fiction. She emphasises the difficulties of making the trip (I knew USSR/Russia had suffered failures, but I wasn't aware of the sheer scale of disaster they've suffered with Martian probes). In a chapter called 'The Massive Mars problem' she dedicates far more to the oddities of Mars being a lot smaller than Earth in models of how the solar system formed than I've seen before, highlighting the unexpected nature of this planet (and, arguably, our solar system as a whole).

Unfortunately there are some issues too. Apart from the overall lack of structure, there are some chapters that just didn't work for me. I like a nice bit of tangential material as much as the next person, but chapter 2, titled 'The Wolf and the Woodpecker' and spending time on Mars as a god, astrology and even, somehow, the tarot, seemed too far removed from the topic. As the first chapter is very general, then we get this, I was really feeling 'Get on with it!' with the obligatory exclamation mark by the end of chapter 2. There are a few other chapters where there's the opposite problem of too much information. In one, Jenner goes through pretty well every Mars probe that has ever existed at some length. Similarly, the breakdown of the impact of microgravity on the human body, with around 15 sections such as 'Bones' and 'Blood and urine' just piles in too much detail. Its a mass of fact statements with no flow. Sometimes less is more.

Pulling a view of the book together, that lack of structure comes through very strongly. For the reader, 4th Rock lacks any sense of narrative drive. It presents collections of information on different topics, some fascinating, some rather tedious - but even all those exclamation marks aren't enough to keep a constant level of interest. Most readers will learn plenty about Mars they didn't know already, so it's definitely worth a look. But for me, it didn't work as well as it could have.

Review by Brian Clegg

Monday, 3 April 2017

A History of the Solar System - Claudio Vita-Finzi ***

It was interesting to read this book straight after Ali Almossawi's Bad Choices. The topics may be totally different, but they are both small books - in this case a mere 100 pages - at high prices for their size. But there the similarity ends. Where the other title conceals a very small amount of content, Claudio Vita-Finzi packs in a huge amount of information on our best understanding (c 2016) of how the solar system came into being, the constitution of its components, where it extends to and far more. But that packing comes at a price, which I'll return to.

I wasn't clear on first seeing the book if the 'history' in the title referred to a sweep through historical views or what has happened to the solar system through time. In practice it does both, but there's relatively little on early ideas, concentrating mostly on our best present theories. It might be a surprise how fuzzy some of those theories are. For example, we still aren't anywhere near certain how the Earth/Moon system formed - and where I thought it was now clear that Earth's water didn't come from comets, as the hydrogen/deuterium ratio is very different, it turns out that this isn't true of all comets... so maybe some of it came from comets after all.

I'd say the ideal use for this book is as an information resource for a student starting a degree in astronomy or a writer wanting some pointers on the solar system. It's a great fact book. Sadly, what it doesn't manage to do is be a good introduction for the general reader. Vita-Finzi packs in so much by making sentence after sentence pure fact statements - almost bullet points - there is very little narrative flow, making the book no easy read.

It doesn't help the readability that the publisher Springer has a weird publication style where each chapter is treated like an academic paper with an abstract, a DOI number and separate references. Bearing in mind an abstract is supposed to be, erm, abstracted, where do they think it's going to go? And those references are particularly obtrusive, both because they are cited with inline numbering like [42] which really breaks the flow of the text and also because, particularly puzzlingly, those numbers seem to be picked at random - in a chapter I just picked at random, the first reference was number 13, followed by number 41.

The slim volume is printed on expensive glossy paper, which means that the colour images could be high quality, though they are often too small to really impress. Nonetheless, apart from reproducing the common repeated misleading assertion that Giordano Bruno was burned at the stake for saying there were many other worlds revolving around other suns (see this post for detailed reasons why this is incorrect), there is a huge amount of useful information in this book that would benefit anyone who needed more detail than is available in a typical popular science title.

Review by Brian Clegg

Sunday, 2 April 2017

Four Way Interview - Marcus Chown

Image courtesy Sunday Brunch, C4
Marcus Chown graduated from the University of London in 1980 with a first class degree in physics. He also earned a Master of Science in astrophysics from the California Institute of Technology. Currently the cosmology consultant for New Scientist magazine, Chown has written a string of successful popular science books. His latest title is The Ascent of Gravity.

Why science?

Good question. When I was 8, my dad bought me 'Dr H. C. King's Book of Astronomy.' I don't know why he bought me that book. But it caught my imagination. Later, he bought me a small telescope. I used to poke it out of the window of our North London flat and observe the stars above the orange glow of the North Circular Road. I saw Jupiter's moons and the rings of Saturn. I began to realise I was living on a tiny mote of matter lost in a mind-bogglingly huge universe. And it awakened in me a desire, which I have never lost, to find out more and more about where it all came from and our place in it.

Why this book?

I was fascinated by the paradoxes of gravity. It was the first force to be discovered, yet today it is the least understood. It is a 'force' that keeps your feet on the ground yet, according to Einstein, no such force actually exists. It is the weakest force in the everyday world, yet it controls the ate of stars and galaxies and the universe as a whole. Gravity, to steal the words of Winston Churchill, is 'a riddle, wrapped in a mystery, inside an enigma.' And penetrating that enigma promises to answer the biggest questions in science: What is space? What is time? What is the universe? And where did it all come from?

What’s next?

I have just been asked to write a Ladybird book. If you remember, Ladybird used to produce those distinctive books for children on topics like 'Going to the dentist' or 'How aeroplanes work.' Well, Penguin revived the imprint a few years ago. They did a range of spoof books for adults [after copying the idea from an artist they sued for doing the same thing - Ed.], which were a publishing phenomenon. So, they have now decided to do an 'expert' series. Prince Charles did one on climate change. I'm chuffed to be doing one of the big bang.

What’s exciting you at the moment?

The fact that nothing fits. We have a model of the universe in which 95% of its mass is invisible - so-called dark matter and dark energy - and we have no idea what it is. We have two incredibly successful theories of physics - quantum theory and Einstein's theory of gravity - and they appear incompatible. All this makes it an incredible time to be alive. We are on the verge of a revolution in our picture of the universe. Someone is going to pull it all together. Someone whose name will become as well-known as Newton and Einstein. At least, I think so!

Bad Choices - Ali Almossawi ***

At the heart of this little book is a really good concept trying to get out - for me it's what you might call a successful failure. What it's trying to do is great, and being creative about doing so is also great - but creativity goes hand in hand with frequent failure, and I'm afraid there are just too many problems here for me to love this book the way I should.

Let's get the negatives out of the way so we can focus on what's good. It's a very short but expensive book - the 144 small pages have a lot of illustration that conveys very little information. What's left is a text that I comfortably read on an hour-long train ride, yet it's being sold at near £15. The illustrations are genuinely almost all doing nothing at all except adding padding. And though Ali Almossawi's writing style is friendly and laid back, it tends to the condescending. Worst of all, the book doesn't do what it says on the tin.

The subtitle is 'how algorithms can help you think smarter and live happier', implying that this book is going to show you how to make use of algorithms to improve your life. It won't. Almost all the examples it uses (in the form of little stories that try far too hard to be quirky) are totally useless in reality. About the only practically useful one is about sorting books on a shelf (though there is better guidance on that elsewhere - see below). And yet. There is something in this book.

What it really does, if you allow it, is to open up the secrets of what's going on inside a computer - specifically in some of the algorithms used to sort and search, to do lookups with hash tables, to have linked lists that enable you to add items to the middle of an ordered set with a minimum of effort and more. There is definitely a beauty, almost a poetry to this stuff, and Almossawi is usually very good at describing it.

So, it's really not going to do what the cover claims. It won't help you with practical, every day choices and problem solving. If you want a book on practical uses of algorithms in real life, look instead at Algorithms to Live By. And those stories, I'm afraid, for me mostly got in the way rather than made the material more approachable. (Almossawi imagines the reader, when asked 'What's a binary search?' thinking 'Ah, freedom, William Wallace, Eppy Toam*, shirts on a rack.' No, we really won't do that.) But for all that there's a lovely little book on a key aspect of how computers do their jobs lurking in here. I just wish there was more content, and it wasn't so obscured by fluff.

* Yes, his characters really do have names like this.

Review by Brian Clegg

Saturday, 1 April 2017

The Future of the Professions - Richard Susskind and Daniel Susskind ***

We're used to hearing how technology is going to replace the jobs of those doing mechanistic jobs - but this book takes on the impact that technology will have on the professions. 

I've only given the book three stars as it feels rather too much like a textbook (admittedly a well-written textbook), it's fairly repetitious and there's limited coverage of the science and technology behind the move. However this doesn't detract from the fascinating aspects of the book.

One of these is simply addressing the professions at all. According to the authors there's a fair amount of literature on this - but it's stuff us ordinary mortals are unlikely to have seen. A starting point is deciding just what the professions are. The book primarily focuses on the traditional professions such as medicine, accountancy, the law, journalism and religion - though they admit that the concept, essentially one where it is necessary to have specialist knowledge and there is often regulation and/or certification, is now a lot wider. (In practice, though religion gets passing mentions, it's largely sidelined, which is probably sensible in the context.)

The authors' assertion is that these roles can be subject to a kind of production line breakdown of tasks, some parts of which can easily be accommodated by information technology or less qualified individuals. The argument is that not only will this reduce costs where, for example, companies are reluctant to continue paying through the nose for corporate law (bye bye Suits), it also has the potential to open up these services to a much wider clientele that is presently largely excluded or at least has significantly reduced access.

Of course there are plenty of objections (often from those involved in the professions) which the authors largely succeed in knocking out of the way. For example they point out that this move will probably reduce the earnings of many professionals - but as they observe, these roles are not there for the benefit of the professionals but for their clients. Inevitably there is quite a lot of futurology style guesswork here. The authors point out they will often be wrong in detail - but argue convincingly that the professions are going to go through a major upheaval in the next generation.

It's amusing, given the authors' assertion that 'in the professions, knowledge resides in the heads of professionals, in books...', using this as a mark of how out of step the professions are in the internet age... that I should have been reading this in a book, rather than, say, a blog post or electronic magazine article. However this still remains a title of interest to anyone either involved in a profession (traditional or more modern) or interested in the future of the middle class.

Review by Brian Clegg