Probabilistic Numerics, NIPS 2012 workshop [deadline extended!]

Dear Colleagues,

we are happy to announce the following NIPS workshop, for which we are gratefully acknowledging PASCAL’s financial support. If you have seen the announcement elsewhere, please note the recently extended deadline.

NIPS 2012 Workshop on Probabilistic Numerics December 8, 2012 at Lake Tahoe, Nevada, US

Traditionally, machine learning uses numerical algorithms as tools.
But many tasks in numerics can also be interpreted as learning problems.
Some examples:

* How can optimizers model the objective function, and how should they use the model to act?

* How should a quadrature method use observations of the integrand to estimate the integral, and at which points should it collect them?

* Can approximate inference techniques be applied to numerical problems?

Many such issues can be seen as special cases of decision theory, active learning, or reinforcement learning, but numerical tasks present exceptional demands on computational cost and robustness, so standard methods from these fields require modification to be useful.

We invite contribution of recent results in the development and interpretation of numerical analysis methods based on probability theory.
This includes, but is not limited to the areas of optimization, sampling, linear algebra, quadrature and the solution of differential equations.

Submission instructions are available at

Important Dates:
* Submission of extended abstracts: November 2, 2012
* Notification of acceptance: November 23, 2012
* Final versions of accepted papers due: December 1, 2012
* Workshop date: December 8, 2012

Invited Speakers (confirmed):
Persi Diaconis, Stanford University
Matthias Seeger, Ecole Polytechnique Fédérale de Lausanne Mark Girolami, University College London

Philipp Hennig, Max Planck Society, Tübingen Michael Osborne, University of Oxford John Cunningham, Washington University in St. Louis

We are grateful for support from the PASCAL network.