Keywords: causality, computational statistics, machine learning, independence testing.
My work focuses mainly on causal inference: we try to learn causal structures either from purely observational data or from a combination of observational and interventional data. We therefore develop both theory and methodology. Our work relates to areas like high-dimensional statistics, computational statistics or graphical models. It's an exciting research area with lots of open questions!
Most of the publications are also on Google Scholar.
We have written a book on causality that is now being copy-edited and that will appear as open access at MIT Press. A first draft is now available for download.
Jonas Peters, Dominik Janzing, Bernhard Schölkopf: Elements of Causal Inference: Foundations and Learning Algorithms
- Our paper 'Kernel-based Tests for Joint Independence' (first author: Niklas Pfister) got accepted at Journal of Royal Statistical Society, Series B.
- Our JRSS, Series B, paper on invariant prediction is now online (with discussions).
- In August 2016, I have joined the statistics group in the Department of Mathematical Sciences at the University of Copenhagen as an associate professor.
If you think about doing a PhD in causality, please send me an email.
Jonas is associate professor at the Department of Mathematical Sciences at the University of Copenhagen.
He is a member of the Junge Akademie.
Until August 2016, Jonas was a group leader for the causality group at the Empirical Inference group at MPI Tübingen.
Before joining Tübingen, he was a postdoc (Marie Curie fellowship) at the Seminar für Statistik, ETH Zurich (CH).
During his PhD and Postdoc he has been working with Dominik Janzing
and Bernhard Schölkopf at the MPI for Intelligent Systems, Tübingen (GER),
and later with Peter Bühlmann
and Nicolai Meinshausen at ETH Zurich;
his thesis received the ETH medal.
Jonas has been working with Leon Bottou at Microsoft Research (WA, USA),
Martin Wainwright at UC Berkeley (CA, USA)
and Peter Spirtes at CMU (Pittsburgh, USA).
He studied Mathematics in Heidelberg (GER) and in Cambridge (UK).
A full CV can be found here.
ACM Transactions on Intelligent Systems and Technology,
Annals of Statistics,
IEEE Transactions of Pattern Analysis and Machine Intelligence,
IEEE Information Theory,
International Journal of Approximate Reasoning,
Journal of American Statistical Association,
Journal of Artificial Intelligence,
Journal of Causal Inference,
Journal of Machine Learning Research,
Journal of the Royal Statistical Society,
Nature - Scientific Reports,
Statistics and Computing,
Transactions on Intelligent Systems and Technology
AISTATS (2015, area chair),
ICML (2012, 2013, 2014),
IEEE Int. Workshop on ML for Signal Proc. (2012),
NIPS (2011, 2015),
UAI (2012, 2013, 2014, 2015, 2016)
- JSM, Baltimore, USA, 2017
- Massachusetts Institute of Technology, Boston, USA, 2017
- Danish Society of Theoretical Statistics, Copenhagen, Denmark, 2017
- ERCIM, Sevilla, Spain, 2016
- LMU Munich, Germany, 2016
- University of Oxford, UK, 2016
- Max Planck Institute for Biogeochemistry, Jena, Germany, 2015
- Tutorial at the German Conference on Pattern Recognition (GCPR), Aachen
- ISI World Statistics Congress, Rio de Janeiro, Brasil, 2015
- DMV, Hamburg, Germany, 2015
- University of St. Andrews, UK, 2015
- University of Cambridge, UK, 2015
- Microsoft Research, Cambridge, UK, 2015
- Weierstrass Institute, Berlin, Germany, 2015
- University of Los Angeles, 2014
- California Institute of Technology, Pasedena, USA, 2014
- Institute of Functional Genomics, Regensburg, Germany, 2014
- ERCIM, Pisa, Italy, 2014
- IMS Annual Meeting, Sydney, Australia, 2014
- Carnegie Mellon University, Pittsburgh, USA, 2014
- University of Amsterdam, The Netherlands, 2014
- University of Nijmegen, The Netherlands, 2014
- NIPS Workshop on Causality, Lake Tahoe, USA, 2013
- IST Austria, 2012
- Max Planck Institute for Dynamics and Self-Organization, Göttingen, Germany, 2012
- Conference on Uncertainty and Artificial Intelligence, Barcelona, Spain, 2011
- ETH Zurich, Switzerland, 2011
- International Symposium on Quantum Thermodynamics, Stuttgart, Germany, 2010
- Causality Workshop, Schloss Dagstuhl, Germany, 2009
- Microsoft Advertising, Bellevue, USA, 2011
Causality in 4 Steps
- Consider the following problem: we are given data from gene A (or B) and a phenotype. Clearly, both variables are correlated. What is the best prediction for the phenotype given we are deleting gene A (or B), such that its activity becomes zero?
- Causality matters: Intuitively, the optimal prediction should depend on the underlying causal structure:
But then, if we do not accept any form of causal notion, we cannot distinguish between these two cases and our best prediction must be: "I do not know."!
Causal Model: If we want to be able to describe the above situation properly, we need a so-called causal model that (1) models observational data and (2) interventional data (e.g., the distribution that arises after the gene deletion) and that (3) outputs a graph. Functional Causal Models (also called Structural Equation Models) are one class of such models, see the figure on the right. If you are interested in more details, see the script below, for example.
Examples of questions that are studied in this field: How can one compute intervention distributions from the graph and the observational distribution efficiently? What if some of the variables are unobserved? What are nice graphical representations? Under which assumptions can we reconstruct the causal model from the observational distribution ("causal discovery")? What if we are also given data from some of the intervention distributions? Does causal knowledge help in more "classical" tasks in machine learning and statistics?
I have written a script on causality that I am more than happy to receive feedback on. Please note that it is still missing some sections. It can be downloaded here.