Junyang Wang, PhD student of Chris Oates, has just submitted a new paper on the Bayesian solution of ODEs to the arXiv. This draws a line under the debate over whether an exact Bayesian solution for an ODE can exist in general (under a Gaussian prior) so is definitely work a read. Here’s a link.
The paper, Bayesian Probabilistic Numerical Methods has been selected as a winner of the SBSS best student paper competition. As a result I’ll be presenting the paper at JSM in Vancouver, at the end of July.
My new paper with Chris Oates and Mark Girolami has just appeared on the arXiv - interested readers can check it out here.
The proceedings of MaxEnt 2016 have been published, and contain my first printed article - you can find it here.
Chris Oates and I recently presented our new paper, Bayesian Probabilistic Numerical Methods, in back-to-back talks at ICERM’s workshop on probabilistic scientific computation. The two talks were filmed and are available online. Here’s Chris with Part 1 and me with Part 2. All the other talks from the workshop are also available, and would make for a very educational binge-watch - see here.
My new paper with Chris Oates, Tim Sullivan and Mark Girolami is now up on the arxiv - interested readers can check it out here.
The FOMICS Winter School on UQ is in full swing. I spoke yesterday, again about our work on probabilistic meshless methods. The discussion was quite spirited, with lots of interest on the ideas behind the paper and probabilistic numerics in general! The slides are available here.
Chris Oates, Francois-Xavier Briol, Mark Girolami and I will be presenting our recent work on PN for ODEs, PDEs and Integration at MCQMC in Stanford this year. You can see a minisite giving more details on our respective talks on Chris’s website here. Our session is on Thursday, from 3:50-5:50pm PST.
Mark Girolami's recent talk at the Alan Turing Institute is up on YouTube. After motivating Probabilistic Numerics as a discipline Mark goes on to describe our PMM paper and present some new results in the direction of evolutionary PDEs.
My new paper, Probabilistic Meshless Methods for Partial Differential Equations and Bayesian Inverse Problems (PMM for short) has just popped up on arxiv. This is work with Chris Oates, Tim Sullivan and Mark Girolami.