Cside 2018

Thursday 8th November, 2018

Biometrika Fellow, Dr Benn Macdonald organised a statistical inference challenge, where participants were asked to infer the parameters of a differential equations model from simulated data with known ground truth. The competition had four subchallenges:

Model 1) An ordinary differential equation (ODE) model of cardiac excitation.
Model 2) A partial differential equation (PDE) model of the pulmonary blood circulation system.
Model 3) A stochastic differential equation (SDE) model of cell migration in response to chemotaxis, with complete observations.
Model 4) Like model 3, but parameter estimation is more challenging due to partial, incomplete observations.

Details, including a list of the competition winners, can be found here:

The inference challenge was co-funded by Biometrika Trust and EPSRC (via SofTMech) to the tune of £10,000. It was widely advertised as a national training event for early career scientists in both Statistics and Mathematical Biology, and has raised the profile of our School, and SoftMech in particular, in the both research communities.

The data were published on 22nd October, and participants were given 10 days to work on the challenge, with a submission deadline of 1st November. The results were announced on 6th November. A workshop will be held on 26th November where the winners, as well as the second and third-placed participants for each sub-challenge, are invited to present their results and methodological approaches.

Our School has been particularly successful in this competition.

SofTMech Postdoc Agnieszka Borowska and Glasgow PhD student Diana Giurghita jointly won (working as a team) challenges 3 and 4.

Former Glasgow Biometrika PhD student Umberto Noè (now a postdoc at the German Centre for Neurodegenerative Diseases) won challenge 2, with SofTMech PhD students Mihaela Paun and Alan Lazarus (working as a team) coming second.

Congratulations to them all on an excellent performance, and thank you to Benn for organising this highly successful event.