Best books for AI Researchers

Mathematics for Machine Learning

by Marc Peter Deisenroth, A. Aldo Faisal, Cheng Soon Ong · 2020

~370 pages~9 hrs to read

Covers the linear algebra, calculus, probability and optimisation behind machine learning, then applies them to regression, PCA, density estimation and SVMs. Read it first if your maths needs refreshing.

Dive into Deep Learning

by Aston Zhang, Zachary C. Lipton, Mu Li · 2023

Interactive-style textbook that pairs each deep learning idea with runnable code, from MLPs to attention and transformers. A good way to implement the basics yourself early on.

Deep Learning

by Ian Goodfellow, Yoshua Bengio, Aaron Courville · 2016

~800 pages~20 hrs to read

The standard textbook on the maths and methods of deep networks, including regularisation, optimisation and generative models. Read the foundations early and return to specific chapters as your research narrows.

Understanding Deep Learning

by Simon J. D. Prince · 2023

A modern, well-illustrated textbook covering transformers, diffusion models, reinforcement learning and the theory of why deep networks generalise. Read it after the basics to see current architectures explained clearly.

Probabilistic Machine Learning: An Introduction

by Kevin P. Murphy · 2022

~850 pages~21 hrs to read

A broad, up-to-date treatment of machine learning through a probabilistic lens, from linear models to deep learning. Use it as the reference you keep open while reading papers.

Reinforcement Learning: An Introduction

by Richard S. Sutton, Andrew G. Barto · 1998

~550 pages~14 hrs to read

The foundational text on reinforcement learning: bandits, dynamic programming, temporal-difference methods and policy gradients. Read it if your work touches agents, RLHF or sequential decision making.

Pattern Recognition and Machine Learning

by Christopher M. Bishop · 2006

~740 pages~19 hrs to read

A rigorous Bayesian treatment of classical machine learning, including graphical models and kernel methods. Worth working through when you need depth that goes beyond recipes.

The Elements of Statistical Learning

by Trevor Hastie, Robert Tibshirani, Jerome Friedman · 2001

~750 pages~19 hrs to read

The statistical foundations of supervised and unsupervised learning, including regularisation, boosting, trees and model assessment. Read it to ground your experiments in sound evaluation and statistics.

Build a Large Language Model (From Scratch)

by Sebastian Raschka · 2024

Builds a GPT-style model step by step in PyTorch, from tokenisation and attention to pretraining and finetuning. Read it when you want to know why the models behave as they do, not just how to call them.

The Craft of Research

by Wayne C. Booth, Gregory G. Colomb, Joseph M. Williams · 1995

~310 pages~8 hrs to read

Covers finding a research question, building an argument from evidence and presenting it. Read it when you start framing your own projects instead of taking them from a supervisor.

The Scientist's Guide to Writing

by Stephen B. Heard · 2016

~310 pages~8 hrs to read

Practical advice on structuring papers, writing clearly and handling submission and review. Useful before your first paper and every time you lead one.

Artificial Intelligence: A Modern Approach

by Stuart Russell, Peter Norvig · 1995

~1140 pages~29 hrs to read

The standard survey of the field, organised around rational agents: search, planning, knowledge, uncertainty, learning and acting. Read the agent and planning chapters to ground modern LLM agents in decades of prior work.