Jonas Peters

Postal address Dr. Jonas Peters Seminar für Statistik HG G 11 Rämistrasse 101 8092 Zürich Switzerland More information Phone   +41 44 632 5777 E-Mail

Since Feb 2013 I am a Postdoc (Marie Curie fellowship) at the Seminar für Statistik, ETH Zurich. I came here in Feb 2012 in order to finish my PhD under the supervision of Peter Bühlmann. Before that I have been working under the supervision of Dominik Janzing and Bernhard Schölkopf at the MPI for Intelligent Systems, Tübingen. In 2011, I have spent three months with Leon Bottou at Microsoft Research (WA, USA).

My full CV is available here.

Research Interest

I am mainly working on causal inference problems. Especially, I am interested in recovering the causal DAGs from iid data (continuous and discrete) and from single instances of time series. During my PhD I developed identifiability results for restricted Structural Equation Models. In my diploma thesis I developed a method that aims at rediscovering the correct time direction when given a real-world time series and its reversed sample.

Currently, I am looking into high-dimensional problems of causal inference.

Teaching

Lecturing

• Seminar in Statistics, ETH Zurich, spring semester 2013

Assisting

• Foundations of Mathematical Statistics, ETH Zurich, autumn semester 2012
  
• Computational Statistics, ETH Zurich, spring semester 2012
• Introduction to Statistics, University of Heidelberg, summer semester 2006
• Analysis II, University of Heidelberg, winter semester 2005
• Analysis I, University of Heidelberg, summer semester 2005

Deutsche SchuelerAkademie

• Wahrscheinlichkeiten als Sprache, Deutsche SchuelerAkademie 2012
• Woher kommt der Strom?, Deutsche SchuelerAkademie 2011
• It's Magic, Deutsche SchuelerAkademie 2010
• Huete, Naegel, Schuhkartons, Deutsche SchuelerAkademie 2009

Publications

2012

• J. Peters, P. Bühlmann: Identifiability of Gaussian Structural Equation Models with Same Error Variances, ArXiv e-print, arXiv:1205.2536
• J. Peters, D. Janzing, B. Schölkopf: Causal Inference on Time Series using Structural Equation Models, ArXiv e-print, arXiv:1207.5136
• B. Schölkopf, D. Janzing, J. Peters, E.Sgouritsa, K.Zhang, J. M. Mooij: On causal and anticausal learning, ICML 2012, 29th International Conference on Machine Learning, 1255-1262.

2011

• J. Peters, J. M. Mooij, D. Janzing, B. Schölkopf: Identifiability of Causal Graphs using Functional Models, UAI 2011, 27th Conference on Uncertainty in Artificial Intelligence, AUAI Press, USA, 589-598.
• D. Janzing, E. Sgouritsa, O. Stegle, J. Peters, B. Schölkopf: Detecting low-complexity unobserved causes, UAI 2011, 27th Conference on Uncertainty in Artificial Intelligence, AUAI Press, USA, 383-391.
• K. Zhang, J. Peters, D. Janzing, B. Schölkopf: Kernel-based Cond. Independence Test and Application in Causal Discovery, UAI 2011, 27th Conference on Uncertainty in Artificial Intelligence, AUAI Press, USA, 804-813.
• J. Peters, D. Janzing, B. Schölkopf: Causal Inference on Discrete Data using Additive Noise Models, IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 33:2436-2450.

2010

• J. Peters, D. Janzing, B. Schölkopf: Identifying Cause and Effect on Discrete Data using Additive Noise Models, JMLR Workshop and Conference Proceedings Volume 9: AISTATS 2010, 13th International Conference on Artificial Intelligence and Statistics, MIT Press, Cambridge, MA, USA, 597-604.

2009

• D. Janzing, J. Peters, J. M. Mooij, B. Schölkopf: Identifying Confounders Using Additive Noise Models, UAI 2009, 25th Conference on Uncertainty in Artificial Intelligence, AUAI Press, USA, 249-257.
• J. M. Mooij, D. Janzing, J. Peters, B. Schölkopf: Regression by Dependence Minimization and its Application to Causal Inference in Additive Noise Models, ICML 2009, 26th International Conference on Machine Learning, ACM Press, New York, NY, USA, 745-752.
• J. Peters, D. Janzing, A. Gretton, B. Schölkopf: Detecting the Direction of Causal Time Series, ICML 2009, 26th International Conference on Machine Learning, ACM Press, New York, NY, USA, 801-808.

2008

• J. Peters, D. Janzing, A. Gretton, B. Schölkopf: Kernel Methods for Detecting the Direction of Time Series, GfKl 2008, 32nd Annual Conference of the German Classification Society, Springer, Berlin, Germany, 57-66.
• P. Hoyer, D. Janzing, J. M. Mooij, J. Peters, B. Schölkopf: Nonlinear Causal Discovery with Additive Noise Models, Advances in Neural Information Processing Systems 21, 22nd Annual Conference on Neural Information Processing Systems (NIPS 2008), Curran, Red Hook, NY, USA, 689-696.

Theses

• Diploma Thesis: Asymmetries of Time Series under Inverting their Direction, University of Heidelberg, 2008
• PhD Thesis: Restricted Structural Equation Models for Causal Inference, ETH Zurich, 2012, Errata
  

Software

The code for most of these papers is available at the causality homepage from the MPI for Intelligent Systems in Tübingen.

Reviewing

Journals

IEEE Transactions of Pattern Analysis and Machine Intelligence, Neurocomputing, NeuroImage

Conferences

ICONIP 2011, NIPS 2011, ICML 2012, UAI 2012, ICML 2013, UAI 2013

Personal Interest

• Deutsche SchuelerAkademie
• Studentenphilharmonie Tuebingen
• Sinfonieorchester Notabene