Hasso-Plattner-Institut
Prof. Dr. h.c. mult. Hasso Plattner
 

Christopher Hagedorn

Name at Birth: Christopher Schmidt

Research Assistant & PhD Student

Phone:+49 (331) 5509 - 1317
Email:christopher.hagedorn@hpi.de
Address:August-Bebel-Str. 88, 14482 Potsdam
Room:V 2.01
Links:dblp, Google Scholar, ResearchGate, GitHubLinkedIn

 

Research Area: Data-Driven Decision Support

Research

Parallel Execution Strategies for Causal Structure Learning on GPUs

Learning causal relationships from observational data is insightful for researchers in multiple domains. For example, in genetic research, gene regulatory networks can be derived from gene expression data. Real world gene expression datasets are often high-dimensional, resulting in long execution times prohibiting the application of constraint-based Causal Structure Learning (CSL) algorithms in practice.
As part of our research on Data-Driven Causal Inference we investigate the adaptation of existing CSL algorithms to utilize parallel processing capabilities of modern hardware in order to speed-up execution. This fosters the application of CSL in real-world settings, both in industry and in research. On multi-core CPUs, we introduce dynamic load-balancing for parallel execution of CSL algorithms to avoid situations of under- or overutilization of compute resources to effectively reduce runtimes. 
In recent years, GPUs have advanced to become a source for massive parallel processing. By utilizing the thousands of parallel processing units and shared memory in GPUs, we achieve a runtime improvement by up to two orders of magnitude for multivariate normal distributed data.
In a next step, we will generalize our GPU-accelerated algorithm to fit different data distributions. For this purpose, we will develop a definition of tasks for parallel execution independent of the current data distribution. Such tasks support fine-grained parallelism, e.g., to speed-up processing of raw observational data during conditional independence tests.

Teaching Activities

Winter Term 2021/22

  • GPU-Accelerated Causal Structure Learning

Winter Term 2020/21

  • Master's Project: A Benchmark Suite for Causal Inference or “How  to model a complex world?”

Summer Term 2020

Winter Term 2019/20

Summer Term 2019

Winter Term 2018/19

Summer Term 2018

Winter Term 2017/2018

Summer Term 2017

Winter Term 2016/2017

Supervised Master's Theses

  • Load Balancing Causal Structure Learning Algorithms in Heterogeneous Computing Systems
  • Parallel Execution Strategies for the PC Algorithm with an Entropy-Based Conditional Independence Test
  • A Scalable GPU-Accelerated Causal Structure Learning Algorithm

Journal Reviewing

  • Journal of Parallel and Distributed Computing
  • Journal of Intelligent Manufacturing
  • IEEE Transactions on Automation Science and Engineering

Publications

  • 1.
    Huegle, J., Hagedorn, C., Schlosser, R.: A kNN-based Non-Parametric Conditional Independence Test for Mixed Data and Application in Causal Discovery. ECML-PKDD 2023, accepted (2023).
     
  • 2.
    Hagedorn, C., Lange, C., Huegle, J., Schlosser, R.: GPU Acceleration for Information-theoretic Constraint-based Causal Discovery. In: Le, T.D., Liu, L., Kıcıman, E., Triantafyllou, S., and Liu, H. (eds.) Proceedings of The KDD’22 Workshop on Causal Discovery, Proceedings of Machine Learning Research (PMLR) 185. pp. 30–60 (2022).
     
  • 3.
    Hagedorn, C., Huegle, J., Schlosser, R.: Understanding Unforeseen Production Downtimes in Manufacturing Processes using Log Data-driven Causal Reasoning. Journal of Intelligent Manufacturing. 33, 2027–2043 (2022).
     
  • 4.
    Braun, T., Hurdelhey, B., Meier, D., Tsayun, P., Hagedorn, C., Huegle, J., Schlosser, R.: GPUCSL: GPU-Based Library for Causal Structure Learning. ICDM Open Project Forum. pp. 1236–1239 (2022).
     
  • 5.
    Huegle, J., Hagedorn, C., Boehme, L., Poerschke, M., Umland, J., Schlosser, R.: MANM-CS: Data Generation for Benchmarking Causal Structure Learning from Mixed Discrete-Continuous and Nonlinear Data. WHY-21 @ NeurIPS 2021 (2021).
     
  • 6.
    Hagedorn, C., Huegle, J.: Constraint-Based Causal Structure Learning in Multi-GPU Environments. In: Seidl, T., Fromm, M., and Obermeier, S. (eds.) Proceedings of the LWDA 2021 Workshops: FGWM, KDML, FGWI-BIA, and FGIR, Online, September 1-3, 2021. pp. 106–118. CEUR-WS.org (2021).
     
  • 7.
    Huegle, J., Hagedorn, C., Perscheid, M., Plattner, H.: MPCSL - A Modular Pipeline for Causal Structure Learning. Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. pp. 3068–3076. Association for Computing Machinery, New York, NY, USA (2021).
     
  • 8.
    Hagedorn, C., Huegle, J.: GPU-Accelerated Constraint-Based Causal Structure Learning for Discrete Data. Proceedings of the 2021 SIAM International Conference on Data Mining (SDM). pp. 37–45 (2021).
     
  • 9.
    Huegle, J., Hagedorn, C., Uflacker, M.: Unterstützte Fehlerbehebung durch kausales Strukturwissen in Überwachungssystemen der Automobilfertigung. In: Götz, S., Linsbauer, L., Schaefer, I., and Wortmann, A. (eds.) Software Engineering 2021 Satellite Events, Lecture Notes in Informatics (LNI). Gesellschaft für Informatik, Bonn (2021).
     
  • 10.
    Schmidt, C., Huegle, J., Horschig, S., Uflacker, M.: Out-of-Core GPU-Accelerated Causal Structure Learning. Algorithms and Architectures for Parallel Processing. ICA3PP 2019. pp. 89–104. Springer International Publishing (2020).
     
  • 11.
    Huegle, J., Hagedorn, C., Uflacker, M.: How Causal Structural Knowledge Adds Decision-Support in Monitoring of Automotive Body Shop Assembly Lines. In: Bessiere, C. (ed.) Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, IJCAI-20. pp. 5246–5248. International Joint Conferences on Artificial Intelligence Organization (2020).
     
  • 12.
    Schmidt, C., Huegle, J.: Towards a GPU-Accelerated Causal Inference. HPI Future SOC Lab - Proceedings 2017. pp. 187–194 (2020).
     
  • 13.
    Schmidt, C., Huegle, J., Bode, P., Uflacker, M.: Load-Balanced Parallel Constraint-Based Causal Structure Learning on Multi-Core Systems for High-Dimensional Data. SIGKDD Workshop on Causal Discovery. pp. 59–77 (2019).
     
  • 14.
    Schmidt, C., Uflacker, M.: Workload-Driven Data Placement for GPU-Accelerated Database Management Systems. Datenbanksysteme für Business, Technologie und Web BTW 2019, 18. Fachtagung des GI-Fachbereichs ,,Datenbanken und Informationssysteme" (DBIS), 4.-8. März 2019, Rostock, Germany, Workshopband (2019).
     
  • 15.
    Schmidt, C., Huegle, J., Uflacker, M.: Order-independent constraint-based causal structure learning for gaussian distribution models using GPUs. SSDBM ’18 Proceedings of the 30th International Conference on Scientific and Statistical Database Management. pp. 19:1–19:10. ACM, New York, NY, USA (2018).
     
  • 16.
    Schmidt, C., Dreseler, M., Akin, B., Roy, A.: A Case for Hardware-Supported Sub-Cache Line Accesses. Data Management on New Hardware (DaMoN), in conjunction with SIGMOD (2018).
     
  • 17.
    Schwarz, C., Schmidt, C.: Interactive Product Cost Simulation on Coprocessors. HPI Future SOC Lab: Proceedings 2015. pp. 103–107 (2017).
     
  • 18.
    Schwarz, C., Schmidt, C., Hopstock, M., Sinzig, W., Plattner, H.: Efficient Calculation and Simulation of Product Cost Leveraging In-Memory Technology and Coprocessors. The Sixth International Conference on Business Intelligence and Technology (BUSTECH 2016) (2016).