We address open challenges in the context of causal structure learning by improvements in both the application of statistical and probabilistic concepts, and GPU-based acceleration to support a utilization in a real-world context.

The knowledge of the causal structures within complex systems is crucial to many domains. For example in manufacturing, where the knowledge of causal relationships constitutes the basis of error avoidance in a production processes. While domain experts within the company have enough expertise to identify the most common relationships, they will require support in the context of both, an increasing amount of observational data and the complexity of large systems. This gap can be closed by algorithms of causal structure learning that derive the underlying causal relationships from observational data, e.g., error messages and inline measurements of a production process.

The questions that motivate most data analysis are not associational but causal in nature. What are the causes and what are the effects of events under observation? Nevertheless, the statistical methods commonly used today to answer those questions are of associational nature. But “Correlation does not imply causation!”, and misinterpretation often results in incorrect deduction. 

In the recent years, causality has grown from a nebulous concept into a mathematical theory. This is largely due to the work of Judea Pearl (2009) who has received the 2011 Turing Award for “fundamental contributions to artificial intelligence through the development of a calculus for probabilistic and causal reasoning”.

Our research in the context of data-driven causal inference concentrates on several workstreams that combined aim to allow an efficient application in real-world scenarios.

Causal Structure Learning:

• Evaluation of real-world scenarios with respect to data characterization and opportunities of CSL
• Introduction of information theoretic consistent dependence measures
• Allow for more flexible learning algorithms in the context heterogeneous data characteristics

Hard- and Software Acceleration:

• Engineering of a modular IT system that enables the application of machine learning techniques in the context of causal inference
• Application of Graphics Processing Units (GPU) to develop efficient implementations
• Examination of integration into In-Memory Databases

Together with different cooperation partners we examine how causal inference can be applied in a real-world context.

Manufacturing:

• Identification of relevant involved factors in an industrial manufacturing process
• Derivation of structural causal graphs in complex production processes, e.g., Bachelor Project: Application of Causal Inference in Automotive Production

Medicine:

• Application of causal inference in the context of gene expression data
• Incorporation of genetic variants and gene expression data to detect causal relationships

Summer Term 2020

• Lecture: Causal Inference - Theory and Applications in Enterprise Computing

Winter Term 2019/20

• Master Project: Causal Reasoning on Enterprise Data

Summer Term 2019

• Lecture: Causal Inference - Theory and Applications in Enterprise Computing

Winter Term 2018/19

• Master Project: Design and Implementation of a Causal Inference Pipeline

Summer Term 2018

• Bachelor Project: Application of Causal Inference in Automotive Production
• Lecture: Causal Inference – Theory and Applications

Winter Term 2017/18

• Bachelor Project: Application of Causal Inference in Automotive Production

Johannes Huegle, Christopher Schmidt, Dr. Matthias Uflacker