Radford M. Neal

Timeline

1. 1977–1980

   Degrees at Calgary

   He earns a BSc in computer science at the University of Calgary in 1977 and an MSc in 1980, with David Hill advising his master's thesis.

2. 1989–1994

   A doctorate on Bayesian neural networks

   He completes a computer science PhD at the University of Toronto under Geoffrey Hinton. His thesis applies hybrid Monte Carlo to Bayesian neural-network models.

3. 1995–2016

   Professor and research chair

   He serves as a professor in Toronto's Statistical Sciences and Computer Science departments from 1995 to 2016. He holds the Canada Research Chair in Statistics and Machine Learning from 2003 to 2016.

4. 1995

   The wake-sleep algorithm

   With Geoffrey Hinton, Peter Dayan, and Brendan Frey, he coauthors the 1995 Science paper on an unsupervised learning algorithm for multilayer networks of stochastic neurons.

5. 1996

   Bayesian neural-network learning

   He publishes a revised version of his thesis as Bayesian Learning for Neural Networks in 1996.

6. 2000–2003

   Adaptive slice sampling

   He describes slice sampling in a 2000 paper as a method that adapts MCMC updates to the sampled distribution. He publishes the paper in Annals of Statistics in 2003.

7. 2011

   Hamiltonian dynamics for MCMC

   His 2011 chapter “MCMC Using Hamiltonian Dynamics” explains how Hamiltonian dynamics can create distant Metropolis proposals and reduce slow random-walk exploration.

8. As of October 2026

   Professor emeritus at Toronto

   He is a professor emeritus at the University of Toronto, with interests in Bayesian modeling and Markov chain Monte Carlo. In 2026, he submits a paper that analyzes Newcomb's problem through concrete prediction scenarios.