The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World

A Must Read for Anyone Interested in What Our Digital Future Looks Like
A word of warning: I’m not a machine learning practitioner and not even mathematically inclined. I am merely an average person with a (reasonable) degree of culture (I suppose I classify as a sociologist) looking out to understand the world around me a little better.Yet I loved this book, it opened my eyes on a world I suspected existed somewhere in the depths of my computer or when I did a Google search but that I never understood. If you are afraid to be engulfed in equations, don’t be. The author prefers to use allegories and is very good at giving simple explanations, making everything (almost) crystal clear.What you get is a comprehensive overview of where machine learning is going. This is hugely important, considering the sometimes disturbing news you get in the media – not just from Snowden but also from “neurotechnologists” who guide political campaigns by embedding cameras in ads, cameras that detect the viewers’ reactions and then adjust the campaign in function of these reactions.Perhaps my only quarrel with this book – but it does not detract from its excellence and the five stars I gave it – is the philosophical position taken by the author. I don’t quite share his optimism about our robotic future. And this for 2 reasons:One, the connection between learning algorithms and data. Domingos does note at the outset: “…the more data they [the learners] have, the more intricate the algorithms can be.” So without “big data”, you don’t get “good” learning algorithms – the “bigger” the data, the better the algorithm. Or as Domingos writes, “the more data we have, the more we can learn”. True enough. But what about the quality and size of the big data? What if the data is error-filled and we’re not aware of it? Can machine learning be aware of something their human masters are not? On what basis? The algorithms are taught to deal with an imperfect world and draw the most likely conclusions, where “most likely” is highly subjective. Or at least, it is “highly subjective” in my opinion and I realize that it is only one opinion and not necessarily one shared by the author.Two, the connection between you as you are and an expanded “digital you”: a “model” of who you are, your work experience, your tastes, a model that the algorithm has learned – and thanks to algorithms, you will be able to do more things and faster than ever before. Domingos describes an extraordinary future where, for example, in your LinkedIn account, “you’ll immediately interview for every job in the universe that remotely fits your parameters (profession, location, pay, etc) LinkedIn will respond on the spot with a ranked list of the best prospects, and out of those, you’ll pick the first company that you want to have a chat with. Same with dating: your model will go on millions of dates so you don’t have to…”Big time-saver obviously. But it comes at a cost: you have to give your parameters (stuff about your work, your likes and dislikes) to the algorithms. Ye who go digital, leave behind any notion of privacy…So users of this algorithmic landscape will have few shaded areas in which to hide.But for me, privacy concerns aside (and I’m not that concerned, I have nothing to hide), there is yet another matter that I find more worrisome. Domingos (on p.283) tells us he is confident that the future can only get better: “In fact, it’s the systems that have a slight edge in serving us better that will, generation after generation, multiply and take over the gene pool.” Then he adds: “Of course, if we’re so foolish as to deliberately program a computer to put itself above us, then maybe we’ll get what we deserve”.Indeed. That is precisely my worry: can we be sure that there won’t be a Doctor Evil who will do so? Can we be sure that there won’t be an Apprentice Sorcerer who might end up doing this even if unintentionally? A mistake can happen…So yes, I’m deeply worried about this algorithmic future and I don’t share the author’s unflagging optimism. But that doesn’t mean it’s a book you shouldn’t read. On the contrary, it’s a must read precisely because it raises all these fundamental questions about the future of humanity. And it is an easy read even for the non-cognoscenti like myself, I highly recommend the introductory chapters, and if you get bogged down in the middle with all the technicalities, no fear, skip to the last two chapters, they are well worth reading!

Thorough survey of trends in machine learning since I got my M.S. in C.S. in 1989
Part of the reason I bought this book was to see if I can still understand current issues in computer science, 30 years after I earned my M.S. in C.S. Back then the hot book in the field was The Soul of a New Machine, by Tracy Kidder, which I very much enjoyed. But time moves on, and some issues that were hard then, such as the four color problem, have been solved since. Artificial Intelligence, on the other hand, has not. Back in 1989, I fully expected some computer to have passed the Turing Test by now, but doing so has proved more difficult than then expected, although computers now routinely trounce human experts in even the most complex games, such as Chess and Go.Thus, I chose this book both to see why we aren’t there yet, and to see if I can still even follow the current arguments.Like The Soul of a New Machine, this book is written to both be accessible to a literate general audience and rigorous enough for specialists. In my opinion, it entirely succeeds, although my son is better than me at math, so I skimmed some of the math-heavy discussions, as my interest is more in the philosophical issues and predictions about coming developments. In that, I was completely satisfied, as just when I was thinking of docking a star because the discussion was getting too deep into the weeds of particular implementations of machine learning, the author returned to a very interesting discussion of what it all means for our current and future society, complete with useful predictions about what jobs are still likely to be around in twenty years, and the inevitability of voters eventually approving a guaranteed basic income for those no longer able to be employed due to robots and computers doing ever more of the work previously done by humans.I was glad to see the author also explored the dark side of all this – the potential for it to control people rather than to free them, and his suggestions on how to make sure the changes achieve good for all rather than for only a few. I was particularly interested in his conclusion that we needn’t worry about machines ruling over us because even the best machine learners lack will, doing only what they have been programmed to do. I also like his idea that although machine learning may be the next step in evolution, it will be one taken with rather than instead of us, with humans branching out along with computers in ways not predictable yet, but likely to be beneficial.Overall, a very good and very hopeful read, and comforting to still be able to follow along after a full career since grad school and years of retirement.Definitely recommended!

100% recommended

I have been in AI for more than 40 years. I know, I know, I am old, and I have seen just about everything in this area of science. My 1985 Ph.D. was one of the first in the area of Associative Memories (a field related to earlier AI ideas). That said, this book has quickly become my favorite! So concise, so clear and clever! If you want to understand where we are in the AI research these days, READ THIS BOOK! Read it and then take Pedro Domingo’s COURSERA video course! I do not want to spoil your enjoyment by telling you much about the Master Algorithm, but what I can tell you is that it reads like Da Vinci Code! It is full of humor and intrigue. Wonderful book from a rising star in machine learning!