Bernoulli Lecture - Efficient stochastic numerical methods for modelling and learning


Event details

Date 13.06.2019
Hour 17:1518:15
Speaker Benedikt Leimkuhler (University of Edinburgh)
Category Conferences - Seminars

Stochastic differential equations are fundamental tools of modern mathematical modelling; they describe everything from the performance of financial instruments to atomic motion to the complex motion of fish schools or bird flocks. In recent years they have also been used increasingly as tools for the parameterization of artificial models such as neural networks in their application to analysis of data sets. Given their great relevance, it is obviously of crucial importance to design numerical methods having high efficiency (i.e. accuracy and stability) for solving stochastic systems. I will describe a number of examples of stochastic differential equations arising in applications, while also presenting a few principles for the construction of high quality numerical methods for their computational treatment. Part of the Semester : Multi-scale Mathematical Modelling and Coarse-grain Computational Chemistry


Practical information

  • General public
  • Free


  • Sara Bonella (CECAM at EPFL), Carsten Hartmann (BTU Cottbus-Senftenberg), Simon Huppert (Sorbonne Université, Pierre - Marie Curie)


  • Valérie Krier

Event broadcasted in