James Ritchie
I am a machine learning researcher and data scientist based in Edinburgh. I was a postgraduate research student at the Centre for Doctoral Training in Data Science, part of the School of Informatics at the University of Edinburgh.
I am interested in Bayesian approaches to machine learning and statistical inference. For my PhD I researched methods for using Bayesian inference and deep learning for scientific applications that have traditionally presented challenges for probabilistic modelling, as part of Iain Murray’s research group. Applications I have worked on include physiological modelling and density estimation for noisy astronomical datasets.
Before moving to Edinburgh I worked as a data scientist and software engineer in London. For my undergraduate degree I studied engineering at the University of Cambridge.
Publications
-
Bayesian Inference for Challenging Scientific Models
James A. Ritchie.
PhD Thesis, University of Edinburgh, 2023
[Edinburgh Research Archive] [PDF] -
Density Deconvolution with Normalizing Flows
Tim Dockhorn*, James A. Ritchie*, Yaoliang Yu, Iain Murray.
*Equal Contribution
Invertible Neural Networks, Normalizing Flows, and Explicit Likelihood Models Workshop, ICML 2020
[arXiV] [PDF] [Virtual Poster Talk] [Code] -
Scalable Extreme Deconvolution
James A. Ritchie, Iain Murray.
Machine Learning and the Physical Sciences Workshop, NeurIPS 2019
[arXiV] [PDF] [Code]
Talks
Counting Coffee Cups and Photons
A remote talk I gave at the ANC workshop introducing some preliminary work on the problem of inferring photon counts and energies from CCD sensors used in astronomy.
[Slides]Density Estimation with Noisy Data
An introduction and background to our work on Scalable Extreme Deconvolution, plus some thoughts on next steps.
[Slides]Cside 2018
We won one of the categories in the Cside 2018 competition, inferring the parameters of an ordinary differential equation model of the cardiac action potential. I gave a talk about our solution at the associated conference.
[Slides]