Jonas Peters
Professor of Statistics
Department of Mathematics, ETH Zurich
Ramistrasse 101, 8092 Zurich, Switzerland
eMail: firstname.lastname@stat.math.ethz.ch
I was flying too much. My last flight was on 30th January 2020. So far, it works out well and I wish I had stopped earlier or at least as early as many others. Let us help each other to avoid future climate change counterfactuals.

 

 

Research

Keywords: causality, machine learning, Earth system science, distribution generalization, policy learning, independence testing.

For my publications, click on `Research' above. Most of the publications are also on Google Scholar.

See here for a short video introducing my research topic (which was produced when I started as a professor at ETH).

 

 

Supervising theses (Master, Bachelor, Semester)

Please always add a CV and your transcripts when asking for supervision of a thesis. Also, we receive a lot of requests for supervising theses; my apologies that I cannot agree to all of such requests.

 

 

Open Positions

There may be an open position announced later in 2024. All details can be discussed during the application process. My apologies that I cannot answer to all individual emails before the deadline.

 

 

Current PhD Students and Postdocs

 

 

Community Service

I am a member of the European Regional Committee (ERC) of the Bernoulli Society. Currently, I am AEing for IEEE Transactions on Pattern Analysis and Machine Intelligence (since Jan 2021), the Journal of Causal Inference (since Jan 2021), and the ACM/IMS Journal of Data Science (since 2023). (In the past, I have also AE'd for Annals of Statistics, the Journal of the American Statistical Association and SIAM Journal on Mathematics of Data Science.) My apologies that I have to decline most of the additional review requests.

 

 

CV

Jonas is interested in using different types of data to predict the effect of interventions and to build statistical methods that are robust with respect to distributional shifts. He seeks to combine theory and methodology and tries to let real world applications guide his research. His work relates to areas such as causal inference, distribution generalization, dynamical systems, policy learning, graphical models, and independence testing. Since 2023, Jonas is professor in statistics at ETH Zurich. Previously, he has been a professor at the Department of Mathematical Sciences at the University of Copenhagen and a group leader at the Max-Planck-Institute for Intelligent Systems in Tuebingen. He studied Mathematics at the University of Heidelberg and the University of Cambridge and obtained his PhD jointly from MPI and ETH.

My full CV is available here (version: February 2024).

 

 

Scholarships and Awards

Test of Time Award (runner-up) at ICML (with B. Scholkopf, D. Janzing, E. Sgouritsa, K. Zhang, and J. Mooij, 2022), Silver Medal of the Royal Danish Academy of Sciences and Letters (2021), COPSS Leadership Academy, awarded by the Committee of Presidents of Statistical Societies (2021), Guy Medal in Bronze, awarded by the Royal Statistical Society (2019), ASA Causality in Statistics Education Award (with D. Janzing and B. Sch\"olkopf, 2018), Teacher of the year at SCIENCE, University of Copenhagen (2018), Read paper to the Royal Statistical Society, London (with P. B\"uhlmann and N. Meinshausen, 2016), Member of the Junge Akademie (2016--2021; board member 2017--2019), Marie Curie fellowship (2013--2015), ETH medal for an outstanding PhD thesis (2013), Scholarhsip of the Studienstiftung des deutschen Volkes (2004--2008), UNWIN prize and election to scholar (Downing College, University of Cambridge, 2007), European Excellence Programme (DAAD, 2006--2007), Kurt-Hahn-Trust (2006--2007), Holderlin Programme (Allianz, 2006--2007), Deutsche SchulerAkademie (2001).

 

 

Outreach and Popular Science

 

 

Book on Mathematical Games

We have written a book on mathematical games that will appear at MIT Press.

Jonas Peters, Nicolai Meinshausen: The Raven's Hat: Fallen Pictures, Rising Sequences, and Other Mathematical Games

Jim Stein has asked us a few questions about our book `The Raven's Hat'. Click here if you are interested in the interview.

Link to MIT Press

 

 

Book on Causality

We have written a book on causality that has appeared as open access at MIT Press. In July 2018, it was awarded the ASA causality in statistics education award.

Jonas Peters, Dominik Janzing, Bernhard Schölkopf: Elements of Causal Inference: Foundations and Learning Algorithms

Link to bibtex

Link to MIT Press

 

The pdf can be downloaded for free from the MIT Press website (look for "This is an open access title" on the left-hand side).

 

 

Alumnae/i (and most recent position in academia if applicable)

 

 

Memberships

Bernoulli Society, Danish Society for Theoretical Statistics, IMS, ISI (elected), Royal Statistical Society

 

 

The photo was made by Jim Hoyer.

Causality in 4 Steps

  1. 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?
      

     

  2. 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."!

     

  3. 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.

     

  4. 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?

 

 

Other

ZOS Zurich

Kammermusikkreis Unterwachingen

Deutsche SchülerAkademie