Author of PRML, heads Microsoft AI research UK
Christopher Bishop
Profile
Christopher Bishop is the closest thing machine learning has to a textbook-writing institution. His 2006 book Pattern Recognition and Machine Learning — universally known as PRML — became the de facto graduate-level introduction to ML for an entire generation of researchers. If you’ve ever seen someone reference “Bishop” without a first name in an ML paper, this is who they meant. The book’s mathematical rigor — Bayesian throughout, no hand-waving — is what made it stick.
He’s a Microsoft Technical Fellow and the founder and Director of Microsoft Research AI for Science, based in Cambridge UK. He joined Microsoft Research in 1997, ran the Cambridge lab from 2015 to 2022, then spun up AI4Science to focus the org on what he believes is AI’s most important application: accelerating the natural sciences. His group has shipped serious work — MatterGen, a diffusion model that designs novel inorganic materials from desired properties, and MatterSim, which screens them for stability — pushing materials discovery away from trial-and-error.
In 2023 he published Deep Learning: Foundations and Concepts, co-authored with his son Hugh Bishop, as a successor to PRML for the transformer era. It became Springer-Nature’s top-selling book of both 2024 and 2025. For developers crossing over from systems work into ML, it’s the cleanest single source on the math behind modern architectures — attention, diffusion, normalizing flows — without the apologetic tone most popular books adopt.
He’s also Honorary Professor at the University of Edinburgh, a Fellow of Darwin College Cambridge, and a Fellow of the Royal Society. In 2008 he gave the Royal Institution Christmas Lectures, which is the closest thing the UK has to a science-communication knighthood. If you want a working definition of “respected elder of ML,” this is it.
Books
Deep Learning: Foundations and Concepts The 2023 successor to PRML, co-written with his son Hugh — covers transformers, diffusion, and modern architectures with the same Bayesian-rigorous treatment that made PRML a classic. Pattern Recognition and Machine Learning The 2006 textbook that trained a generation of ML researchers — comprehensive, mathematically serious, and unapologetically Bayesian.Videos
Key Articles & Papers
MatterGen: A Generative Model for Inorganic Materials Design The Revolution in Scientific DiscoverySpotify Podcasts
