Peter Wirnsberger


Hi, I am a Staff Research Scientist at Isomorphic Labs, where I develop AI to revolutionise drug discovery. I previously worked at Google DeepMind on problems at the intersection of ML, physics and chemistry.

My research interests include generative modelling, learning simulation and enhanced sampling with large-scale applications in statistical mechanics and fluid dynamics.

Updates

[Paper] GraphCast: Learning skillful medium-range global weather forecasting

Published on the arXiv

[Paper] Gibbs free energies via isobaric-isothermal flows

Published on the arXiv


Lecture at the Erwin Schrödinger Institute

The material is intended for a 3hr lecture followed by a 3hr tutorial.


[Paper] MultiScale MeshGraphNets

Published at IMCL 2022, AI4Science Workshop, arXiv


[Paper] Normalizing flows for atomic solids

Published in Machine Learning: Science and Technology with code and trained models available on github.


[Paper] Targeted free energy estimation via learned mappings

Selected as a featured article by JCP.


Promotion to Senior Research Scientist


2018 Highly Commended Thesis Prize

I received this prize from the Department of Chemistry, University of Cambridge, for my PhD Thesis.


[Paper] Microscopic analysis of thermo-orientation in systems of off-centre Lennard-Jones particles

arXiv | JCP


Dr Peter Wirnsberger

Here is a link to my PhD Thesis.


New position at DeepMind in London

I am excited to share that I will be joining DeepMind as a Research Scientist.


[Paper] Theoretical Prediction of Thermal Polarization

arXiv | PRL


I won a Microsoft Azure Research Award

Microsoft will support my research for an entire year by providing me with access to their cloud computing service Azure.


We won the Microsoft Hackathon

Our team is one of the global winners of this year's Microsoft Hackathon with more than 18,000 participants worldwide.


ML internship at Microsoft

I am very excited about starting my 3-month research internship at Microsoft Research Cambridge today.


[Paper] Numerical Evidence for Thermally Induced Monopoles

arXiv | PNAS


[Paper] Non-equilibrium simulations of thermally induced electric fields in water

arXiv | JCP


[Paper] An enhanced version of the heat exchange algorithm with excellent energy conservation properties

Selected as a 2015 Editors' Choice article.

arXiv | JCP