Data-Driven Decision Support

Causal structure learning (CSL), i.e., the derivation of the causal graphical model representing the functional relationships between variables of a system from observational data, is crucial to many domains. However, while current methods of CSL assume the same data type or a specific family of functional relationships, most real-world scenarios incorporate heterogeneous data, e.g., mixed data. Moreover, as the underlying functional causal model and corresponding consistent model-based dependence measures are mostly unknown in advance the selection of inappropriate methods for CSL yields incorrect results.

As part of our research on Data-Driven Causal Inference we address these challenges through the introduction of information theoretic statistical methods and the development of a modular pipeline for experimental evaluation of CSL from observational data. We demonstrate the applicability of our approaches on both synthetic and real-world settings with regard to the consistency of the estimated underlying causal structures.

Keywords: Causal Structure Learning, Causal Inference, Machine Learning in Real-World Scenarios, Computational Statistics

• Data-Driven Causal Inference
  • Information-Theoretic Causal Structure Learning from Mixed and Nonlinear Data
  • MPCSL - A Modular Pipeline for Causal Structure Learning
• Application Scenarios with Cooperation Partners
  • Mechanical Engineering
  • Automotive Production

Lectures:

• Causal Inference - Theory and Applications in Enterprise Computing (SS18, SS19, SS20)
• Trends and Concepts in the Software Industry I (SS17)
• Trends and Concepts in the Software Industry II (WS16/17)

Seminars:

• Trends in Bioinformatics (WS17/18, WS18/19)

Master Projects:

• GPU-Accelerated Causal Structure Learning (WS21/22)
• A Benchmark Suite for Causal Inference or “How to model a complex world?” (WS20/21)
• Causal Reasoning on Enterprise Data (WS19/20)
• Design and Implementation of a Causal Inference Pipeline (WS18/19)

Bachelor Projects:

• Online Marketplace Simulation: A Testbed for Self-Learning Agents (WS21/22,SS22)
• Application of Causal Inference in Automotive Production (WS17/18, SS18)
• Machine Learning for Sales Order Fulfillment (WS16/17, SS17)

(Co-) Supervised Master's Theses:

• Causal Structure Learning for Car Manufacturing using Background Knowledge (Tobias Nack)
• Parallel Execution Strategies for Constraint-Based Causal Structure Learning with an Entropy-Based Conditional Independence Test (Constantin Lange)
• Constraint-based Causal Structure Learning of Gene Regulatory Networks (Philipp Bode)
• An Entropy-Based Conditional Independence Test for Causal Structure Learning from Heterogeneous Data (Daniel Thevessen)
• A Scalable GPU-Accelerated Causal Structure Learning Algorithm (Siegfried Horschig)

Master's Degree in Mathematics, Heidelberg University, 2017, Focus: Statistics and Probability Theory, Thesis: "Asymptotics for Directed Interaction Networks based on Counting Process Theory"

Bachelor's Degree in Mathematics, Heidelberg University, 2014, Focus: Statistics and Probability Theory, Thesis "Pricing of Derivatives in the Binomial Model"

• "Causal Discovery in Practice: Tooling & Non-Parametric Conditional Independence Testing for Causal Discovery" European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD) - PhD Forum, 2023.
• "Causal structure learning for heterogeneous data characteristics of real-world scenarios", European Causal Inference Meeting (EuroCIM), 2021.
• "An Information-Theoretic Approach on Causal Structure Learning within Heterogeneous Data Characteristics of Real-World Scenarios", SIAM International Conference on Data Mining (SDM) - Doctoral Forum, 2021.
• "Probabilistic Machine Learning - Introduction to Causal Inference", SAP Innovation Series, 2017.