If I could go back in time to the start of my Ph.D., these are a few of the books I would make myself read. This list is both personalized to my interests and non-exhaustive, but I think each one of them has useful lessons for any AI/ML Ph.D. student.
These are a few books on technical topics that have useful lessons for machine learning researchers.
This book is worthwhile even if you’re not especially interested in the neuroscience. It provides a nice perspective on the role of mathematical models in understanding complex systems, as well as some history of machine learning as a discipline.
If you intend to apply machine learning to medicine or molecular biology, this is an excellent primer on genetics. Even if you don’t care about biology at all, I still think you should read this book. You’ll get a humbling sense of how simple machine learning is (both conceptually and technologically) compared to the complex and difficult questions being tackled by scientists in other disciplines. You’ll also learn the story of how another field dealt with the unexpected arrival of powerful and potentially dangerous technologies like genome editing.
This is a textbook about how humans form categories and reason about their relationships. This is not a light read, but I really like it and return to it often. You’ll be pushed to think more deeply about how we often simplify the world for our machine learning problems.
Writing is a hugely important part of a Ph.D., and one for which we are very poorly trained. These books can offer some guidance and inspiration, though there is really no substitute for regular practice.
This is a very readable book of nonfiction writing tips. While the whole book is worth reading, there is a great chapter specifically focused on writing about science and technology. That chapter opens with the idea that “writing is not a special language owned by the English teacher. Writing is thinking on paper. Anyone who thinks clearly can write clearly, about anything at all. Science, demystified, is just another nonfiction subject. Writing, demystified, is just another way for scientists to transmit what they know.”
This book is less about specific writing tips than it is about the process and practice of writing. I was struck by how much of the advice in this book rings just as true for finishing a Ph.D. as it does for finishing a novel: “E.L. Doctorow once said that ‘writing a novel is like driving a car at night. You can see only as far as your headlights, but you can make the whole trip that way.’ You don’t have to see where you’re going, you don’t have to see your destination or everything you will pass along the way. You just have to see two or three feet ahead of you.”
Doing a Ph.D. is largely about managing your brain and body. These books are helpful for understanding why it’s so important and how to do it better.
This book makes a compelling scientific argument that prioritizing sleep is the foundation for being happy, healthy, and productive. Being sleep-deprived for your entire Ph.D. might be rather common, but it turns out that it won’t do you many favors in the short term or the long term.
This is a superstition-free book about mindfulness meditation. It’s short, well-written, and provides some useful tools for managing negative emotions and stressful situations. This is important because your ability to handle frequent rejections and setbacks will be a major factor in how well your Ph.D. goes.
This book takes a thoroughly practical approach to the problem of using your time and resources well. The basic point is that you won’t be able to do most of the things you might want to do in life (or during your Ph.D.). This means you need to aggressively prioritize the important things, and then you need to get comfortable with the fact that you simply won’t get around to the rest. I think this book will resonate with anyone who has to engage with the frantic pace of machine learning research.