Algorithms to live by, by Brian Christian and Tom Griffiths｜Book Review｜
Hello Readers! This week I challenged myself by reading a book in a field I’m not very familiar with…Computer Science! We are discussing Algorithms to live by, by Brian Christian and Tom Griffiths.
This book is about the computer science of human decisions.
Here is my review!
It’s a term I’d never heard of, then again, I was unfamiliar with many terms in this book! Essentially, Overfitting means over-compensating. Christian and Griffiths explain that people will build whatever the leader is measuring. Now, this can have considerable impact in business. Today’s business culture is data-driven and focused on hitting key metrics. We measure ROI, click-through rate, subscribers, likes, shares, it goes on and on! Whatever you, your team, or your company recognize as the “key metric”, is what people will focus on. Sounds good right? It allows us to measure how well we are performing assuming we choose appropriate metrics. That’s not the full story though. How do we know someone is a top-performer and not simply hitting a specific metric very well. The nature of overfitting causes employees to maximize their performance in areas that “count”, ignoring other aspects. So what’s the solution, how do we know someone is actually a top-performer?
According to Christian and Griffiths, the answer is Cross Validation. In technical terms, Cross Validation is:
a model validation technique for assessing how the results of a statistical analysis will generalize to an independent data set.
For those of you that are like me and don’t really know what that actually means, let me “dumb it down!” Cross Validation involves selecting a key metric and measuring it, but also having one or two additional metrics that you can use to cross-reference and see if an individual is performing well across the board or not. If someone is performing well on the primary metric but scoring poorly on the other two metrics, that is a strong indicator that overfitting is occurring. However, if the primary and secondary metrics are being achieved, this indicates an all-star performance by the employee. It’s important to be aware of the potential for overfitting, otherwise the numbers can fool you!
To try and fail is at least to learn; to fail to try is to suffer the inestimable loss of what might have been.
As a novice in the field of computer science and math, I appreciated the way Christian and Griffiths simplified the rather complex concepts that are discussed. You are given a brief history lesson, explaining how each concept came about and are then shown how these formulas and concepts are implemented in the real world. I enjoyed how the authors made the information relatable and meticulously discussed each topic.
No chapter summaries! If there was ever a book that would have benefited from chapter summaries, it’s this one right here! Perhaps readers that are familiar with this field of study won’t mind, but as a novice, I found it very difficult to keep track of all the information that was presented within a chapter, yet alone the entire book. Chapter summaries would have helped me remember key concepts without having to skim through entire sections to refresh my memory on a topic. Another issue I had with this book, is that I didn’t find the information particularly useful in terms of actually implementing the concepts into my daily life. The information is relevant to real-world settings, and is crucial in how our world functions, however in terms of actionable advice that can improve your life, I didn’t find that information here.
Who should read this book?
It should go without saying that people who love math, data analytics and computer science will enjoy this book. Also, if you love a challenge and want to explore a new area of study I think this book is a great starting point.