Thursday, December 08, 2016

Review: Algorithms to Live By

Algorithms to Live By is actually an interesting book, even if you're a computer scientist. For one thing, even if you're a computer scientist, once in a while it's nice to get a layman's refresher of the breadth of the field. This includes analysis of the Secretary Problem, a review of exponential backoff, and of course, the range of NP-complete problems (though since the book is for a non-technical audience, the term NP-complete is never mentioned)

Some of the applications are fun and entertaining, like the application of scheduling theory to personal time management.  Others are a mere discussion of topics: for instance, over-fitting in machine learning has correspondences in human training programs. Most Googlers would have already had exposure to these algorithms, as well as the quick sampling of auction theory the book covers.

For the non-technical, this book is written in a very clear, non-mathematical fashion. As long as you remember what a polynomial and an exponent with, I don't expect any of the concepts in the book to give you trouble. Since the book opens with the secretary problem, a quick browse of the first chapter will quickly help you determine whether this book is for you. Certainly, for a non-technical audience I find it hard to think of a better book to read if you're not familiar with computer science.


Post a Comment