Max Planck Research Group Leader

Office: 223

Spemannstr. 38

72076 Tübingen

Germany

Spemannstr. 38

72076 Tübingen

Germany

+49 7071 601 572

+49 7071 601 552

I am interested in algorithms that estimate and express uncertainty about the result of imprecise computations. Such imprecision can arise because the computational task is not analytically tractable, because a limited computational budget only allows a partial solution, or because the description of the task is itself imprecise to begin with. Probability measures provide the formal language for the description of such uncertainty. My group and I develop computer algorithms that take in and return probability measures; we call these probabilistic numerical methods.

- For detailed information about our work, please see our group's homepage.
- A general overview of probabilistic numerical methods can be found at http://www.probabilistic-numerics.org.
- I was recently interviewed by Katie Gorman and Ryan Adams for Talking Machines. The interview provides a bit of a general introduction. In a bittersweet coincidence, this episode of the podcast opens with an obituary for my incredible PhD advisor, Sir David MacKay.
- It usually relatively up to date. The list of publications on this webpage, however, is usually more reliable.

- Video of my talk on Bayesian Optimization at the Gaussian Process Summer School 2015
- I organised the Machine Learning Summer School 2015 (videos and slides). During this MLSS, I also did a tutorial on probabilistic numerics
- Videos of my Gaussian Process tutorial at the RNLS summer school at the ETH Zürich are online. This is the most recent and most compact video of my GP tutorial. If you want to learn about GPs in 90 minutes, I recommend watching this one instead of the older ones listed below.
- I organised the Machine Learning Summer School 2013 (videos and slides)
- Videos of my talks at the MLSS are up (slide animations only work in Adobe Reader):
- I spoke on GPs at the Gaussian Process Winter School 2013. Here is a video, and the slides of this talk (slide animations only work in Adobe Reader).

can be found here. Don't trust it to always be up to date. If you need a bio-blurb for your event web-page or a talk introduction, here's a suggestion (sorry if this sounds like grandstanding, I've repeatedly been asked for such a text):

*Philipp Hennig heads the Max Planck Research Group on Probabilistic Numerics at the MPI for Intelligent Systems in Tübingen, Germany. He studied physics in Heidelberg and London, before he moved to Cambridge, UK, where he did his PhD in David MacKay's inference group. Since that time, he is interested in the information content of computations, and mathematical notions of uncertainty for deterministic computation. Together with two colleagues from Oxford and Columbia U, he organized the inaugural workshop on Probabilistic Numerics in 2012, (re-) starting a community effort to provide a rigorous formulation of computation as the collection of information by autonomous, self-consistent agents. His work has been published in the leading venues of machine learning, as well as journals of the applied mathematics community. Together with his group, he has provided novel interpretations of classic numerical algorithms as maximum a-posteriori estimators, and used these results to create new algorithmical tools for machine learning and artificial intelligence.*

Inference Probability Numerical Methods

45 results
(BibTeX)

**Using an Infinite Von Mises-Fisher Mixture Model to Cluster Treatment Beam Directions in External Radiation Therapy **
In pages: 746-751 , (Editors: Draghici, S. , T.M. Khoshgoftaar, V. Palade, W. Pedrycz, M.A. Wani, X. Zhu), IEEE, Piscataway, NJ, USA, Ninth International Conference on Machine Learning and Applications (ICMLA), December 2010 (inproceedings)

**Approximate Inference in Graphical Models**
University of Cambridge, November 2010 (phdthesis)

**Expectation Propagation on the Maximum of Correlated Normal Variables**
Cavendish Laboratory: University of Cambridge, July 2009 (techreport)

**Bayesian Quadratic Reinforcement Learning**
NIPS Workshop on Probabilistic Approaches for Robotics and Control, December 2009 (poster)

**Point-spread functions for backscattered imaging in the scanning electron microscope **
*Journal of Applied Physics *, 102(12):1-8, December 2007 (article)