I am a PhD student at the Max-Planck-Institute for Intelligent Systems where my supervisor is Philipp Hennig. I'm working on probabilistic numerical algorithms for solving ordinary differential equations. This means that I'm curious what happens if you look at numerical algorithms from an inference/estimation perspective. More specifically, I try to answer two questions:

  1. How can we interpret traditional numerical algorithms in an inference framework? What new insights does this give us about existing algorithms?
  2. If we ignore traditional approaches to solving ODEs for the moment, what can we do in an inference framework that has not been possible previously?

Using Gaussian process models we have made some progress with both questions.

News

2016-05-18
I have written a blog post on probabilistic numerics
2016-04-28
I organized the first Probabilistic Numerics PhD students convention in Tübingen
2015-10-01
I had the opportunity to work with Sebastian Nowozin this summer. Stay tuned for our publication
2015-02-01
New website online
2014-09-09
Our paper "Probabilistic ODE solvers with Runge-Kutta means" was accepted for oral presentation at NIPS
2014-05-23
I was awarded a travel award from the MICCAI society

You can also follow me on Twitter.

Curriculum Vitae

You can download my full CV as PDF.

Education

2013 - now
PhD student at the Max-Planck-Institute for Intelligent Systems, Tübingen. Supervisor: Philipp Hennig
2006 - 2013
Diploma degree in computer science at the Eberhard-Karls-University Tübingen (GPA: 1.1)
2010 - 2011
Study abroad program at the University of Maryland, College Park

Professional Experience

2015
Internship at Microsoft Research, Cambridge, UK: extending a probabilistic model for time-of-flight depth imaging
2011
Internship at LOCOM Software: implementing a transport rate calculation tool based on linear programming
2008 - 2010
Research assistantship: implementing a Java-API for linear program solvers
2007 - 2008
Teaching assistant for the introductory computer science class at the Eberhard-Karls-University Tübingen

Awards and Honours

Contact

Michael Schober
Research Group Elementary Intelligence
Department Empirical Inference
Max-Planck-Institute for Intelligent Systems