Posters displayed in this page are ordered by alphabetical order of the speakers' surnames. The posters will be displayed during the coffee breaks, where we encourage participants to inquiry about them to the authors attending the conference.

Computational barriers and condition in the mathematics of information

Alexander James Bastounis (U. Cambridge), Anders Hansen (DAMTP, Cambridge), Verner Vlacic (ETH)

This poster concerns the analysis of computational barriers in a variety of optimisation problems. We have shown that popular methods such as LASSO (used in feature selection), basis pursuit (used in compressed sensing) and linear programming cannot be computed with a computer that can only read approximate input with arbitrarily precise accuracy. There are important links between this work and condition, and this shall be demonstrated and discussed in further detail on the poster.

Homotopy Continuation for 3264 Conics

Paul Breiding (TU Berlin), Bernd Sturmfels (MPI MiS), Sascha Timme (TU Berlin)

I will present a website where the user can compute the 3264 conics tangent to their own 5 conics.

On the condition number of the real polynomial eigenvalue problem with random inputs

Carlos Beltran (Universidad de Cantabria), Khazhgali Kozhasov (TU Braunschweig)

We explicitly compute the average condition number for real polynomial eigenvalues of independently distributed Gaussian matrices.

The SQRA-Discretization for the Fokker-Planck equation

Martin Heida (WIAS), Artur Stephan (HU Berlin)

We present the newly developed SQRA finite volume discretization and prove convergence of order 1 for the solution and the fluxes. We further show that SQRA is asymptotically equivalent with the Scharfetter-Gummel scheme.

A Mathematical Theory of Learning Guided by the Immune System

Indika Rajapakse (University of Michigan), Roger Brockett (Harvard University)

Develop a mapping from the mechanisms used by the human immune system to confront antigens and build a lifetime of immunity to algorithms and memory strategies that can be used for lifelong learning in engineering applications.