Philipp Geiger

Address, contact upon request; web: geiger.onl;
email: upon request, phone: upon request; GitHub, Google Scholar

Summary

Education, research Doctorate in computer science, MSc (eq.) in mathematics; published at ICML, UAI; machine learning, causal inference, time series, multi-agent, economic decisions
Applications, teamwork Python, MySQL; Gaussian processes, RNNs; demand forecasting app; cloud models; cooperating, presenting results with/to researchers, engineers, diverse stakeholders

Experience

04/2017 – present
Postdoc researcher
Max Planck Institute for Intelligent Systems, Tübingen, Germany
  • Leading research project on machine learning for efficient multi-agent facility usage
  • Applying time series analysis: Kalman filtering, exponential smoothing, RNNs
  • Applying game theory: Bayesian games, best-response dynamics; using sensor data
  • In Python, TensorFlow, MySQL: implemented data-driven congestion forecasting app
  • Teamed up with economist, physicists, software engineers; supervised a MSc student
  • Achieved data privacy agreement with work councils after presenting project to them
07/2015 – 10/2015
Research intern
Microsoft Research Ltd., Cambridge, United Kingdom
  • Implemented AI agents in simulations together with engineers under Katja Hofmann

Education

06/2013 – 03/2017
Doctorate in computer science (equivalent to PhD)
Max Planck Institute for Intelligent Systems, Tübingen, and University of Stuttgart, Germany
  • Thesis title: "Causal models for decision making via integrative inference"
  • Grade: magna cum laude/"very good"
  • Supervisors: Bernhard Schölkopf, Dominik Janzing and Marc Toussaint
  • Connected causal models with quasi-experiments, counterfactuals, information theory, reinforcement learning, statistics, economic time series analysis, decision making
  • Used Gaussian process regression for cloud computer debugging; Python, R, Matlab
10/2006 – 12/2012
Diplom in mathematics (equivalent to MSc)
Heidelberg University and Humboldt University of Berlin, Germany
  • Thesis title: "Mutual Information and Gödel Incompleteness"
  • Grade: 1.4 (best score 1.0 of 5.0)/"very good"
  • Specialization: mathematical logic, theoretical computer science; minor: philosophy

Skills

Program-ming
  • Machine learning implementation (Gaussian process regression, ridge regression, neural networks, Kalman filtering, exponential smoothing, k-means clustering) with Python (working knowledge), TensorFlow, R, Matlab, MySQL (basic) in Linux
  • Object-oriented programming with Python (working knowledge), C++ (basic)
Communi-cating
  • Presenting and explaining data, insights and results using PowerPoint, LaTeX, HTML
  • Coordinating with diverse stakeholders and understanding them: customers, manufacturers, researchers, software engineers, work councils and privacy officers
  • Languages: German (native), English (fluent), French (beginner)

Selected publications

Peer-reviewed
  1. Geiger, P., Zhang, K., Gong, M., Janzing, D., & Schölkopf, B. (2015). Causal inference by identification of vector autoregressive processes with hidden components. In Proceedings of the 32nd International Conference on Machine Learning (ICML).
  2. Gong, M., Zhang, K., Schoelkopf, B., Tao, D., & Geiger, P. (2015). Discovering temporal causal relations from subsampled data. In Proceedings of the 32nd International Conference on Machine Learning (ICML).
  3. Geiger, P., Janzing, D., & Schölkopf, B. (2014). Estimating causal effects by bounding confounding. In Proceedings of the 30th Conference on Uncertainty in Artificial Intelligence (UAI).
Preprints
  1. Geiger, P., Winkelmann, J., Proissl, C., Besserve, M., & Schölkopf, B. (2018). Coordination via predictive assistants from a game-theoretic view. ArXiv Preprint ArXiv:1803.06247.
  2. Geiger, P., Carata, L., & Schoelkopf, B. (2016). Causal inference for cloud computing. ArXiv Preprint ArXiv:1603.01581.
Theses
  1. Geiger, P. (2017). Causal models for decision making via integrative inference. PhD thesis.
  2. Geiger, P. (2012). Mutual information and Gödel incompleteness. Diploma thesis.

Supervision, teaching and reviewing

10/2016 – 03/2017
Supervisor
  • Student: Claudius Proissl (University of Stuttgart); research project during MSc
10/2013 – 02/2014
Teaching assistant
University of Tübingen, Germany
  • Lecture "Intelligent Systems I": a first course in machine learning
10/2011 – 04/2012
Teaching assistant
Heidelberg University, Germany
  • Lecture "Computability and Computational Complexity Theory I"
10/2014 – present
Reviewer
  • Conferences: NIPS ('14, '17), ICML ('16, '17), UAI ('16, '17, '18)
  • Journals: ACM TIST, IEEE PAMI, IEEE TKDE, IJDSA

Memberships and awards

09/2015 – 06/2017 Associate Doctoral Fellow of Max Planck ETH Center for Learning Systems
07/2005 Award for outstanding results in physics by German Physical Society (DPG)

References

Prof. Bernhard Schölkopf Max Planck Institute for Intelligent Systems, Tübingen, Germany
Dr. Katja Hofmann Microsoft Research Ltd., Cambridge, United Kingdom
Dr. Wolfgang Merkle Heidelberg University, Germany