Prof. Dr. rer. nat. Marius Lindauer
Leibniz Universität Hannover
Institut für Informationsverarbeitung
Appelstr. 9A
30167 Hannover
Germany
phone: +49 511 762-5301
fax: +49 511 762-5333
office location: room 1306

Mission Statement:

In recent years, AI made impressive results in different applications possible, e.g., in computer vision, natural language processing or game playing. These breakthroughs show how AI will influence and change our daily lives in many fields. With the advent of deep learning and also traditional AI methods, such as AI planning, SAT solving or evolutionary algorithms, a multitude of different techniques are available these days. However, applying these techniques is challenging and even experienced AI developers are faced with several difficult design decisions, such as which algorithms to apply and how to set their corresponding hyperparameters. Unfortunately, the performance and thus the success of AI systems strongly depend on these small but important design decisions. To make AI easy-to-use for more users (even for those without a strong background in AI), we develop automated machine learning (AutoML) methods, which address many different research questions, such as: (i) how to predict the best algorithm for a given input? (ii) how to efficiently search for well-performing hyperparameter settings of an algorithm? (iii) how to efficiently analyze the performance of algorithms and their inputs? Finding solutions to such problems will lead to a democratization of AI which will unleash the true potential of AI.

Short CV:

  • since Oct 2019: Professor for machine learning at the Leibniz University of Hannover
  • 2014-2019: Postdoctoral research fellow at the University of Freiburg with focus on AutoML
  • 2015: PhD in computer science at the University of Potsdam (summa cum laude) with focus on algorithm selection and configuration
  • 2013: Co-Founder of the research network COSEAL (since 2018 advisory board member of COSEAL)
  • 2010: Master of Science in computer science at the University of Potsdam
  • 2008: Bachelor of Science in computer science at the University of Potsdam
  • 2005: High school graduation (Abitur) in Berlin

Selected Awards:

  • 2018: Winner of 2nd AutoML challenge::PAKDD2018 with aad_freibug and PoSH Auto-sklearn
  • 2016: Winner of ChaLearn AutoML challenge "AutoML 5" with aad_freibug and auto-sklearn
  • 2015: Winner of ICON Challenge on algorithm selection with AutoFolio (track: Par10)
  • 2013: Winner of Configurable SAT Solver challenge 2013 with the Potassco team and clasp (tracks: crafted and random)
  • 2012: Winner of SAT Challenge 2012 with the Potassco team and clasp (track: hard combinatorial)
  • 2011: Winner of Answer Set Programming Competition with the Potassco team and claspfolio (track: NP-Problems)
  • 2009: Leopold-von-Buch-Bachelor-Award (Best Bachelor in Natural Sciences 2009 at the University of Potsdam)

Snippet of Research Interests

  • Andre Biedenkapp, H. Furkan Bozkurt, Theresa Eimer, Frank Hutter, Marius Lindauer
    Algorithm Control: Foundation of a New Meta-Algorithmic Framework
    Proceedings of the European Conference on Artificial Intelligence (ECAI), 2020
  • M. Lindauer and J. N. van Rijn and L. Kotthoff
    The Algorithm Selection Competitions 2015 and 2017
    Artificial Intelligence, pp. 86--100, December 2019
  • Marius Lindauer and Frank Hutter
    Best Practices for Scientific Research on Neural Architecture Search
    Arxiv/CoRR, September 2019

Show all publications
  • Conference Contributions
    • A. Biedenkapp and H. F. Bozkurt and F. Hutter and M. Lindauer
      Towards White-box Benchmarks for Algorithm Control
      {IJCAI} 2019 {DSO} Workshop, August 2019
    • L. Fuks and N. Awad and F. Hutter and M. Lindauer
      An Evolution Strategy with Progressive Episode Lengths for Playing Games
      Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI’19), pp. 1234-1240, August 2019
    • M. Feurer and K. Eggensperger and S. Falkner and M. Lindauer and F. Hutter
      Practical Automated Machine Learning for the AutoML Challenge 2018
      ICML 2018 AutoML Workshop, July 2018
  • Book Chapters
    • Hector Mendoza and Aaron Klein and Matthias Feurer and Jost Tobias Springenberg and Matthias Urban and Michael Burkart and Max Dippel and Marius Lindauer and Frank Hutter
      Towards Automatically-Tuned Deep Neural Networks
      AutoML: Methods, Sytems, Challenges, Springer, pp. 141--156, December 2018, edited by Hutter, Frank and Kotthoff, Lars and Vanschoren, Joaquin
  • Technical Report
    • Marius Lindauer and Frank Hutter
      Best Practices for Scientific Research on Neural Architecture Search
      Arxiv/CoRR, September 2019
Other activities

Research homepage:

Publications:

Recently recorded talks:

Open-source projects:

  • SMAC v3: automatic tuning of hyperparameter configurations on any kind of algorithms (mainly based on Bayesian Optimization)
  • AutoPyTorch: automatic hyperparameter optimization and architecture search for deep neural networks
  • CAVE : Configuration Assessment, Visualization and Evaluation
  • Auto-Sklearn: automated machine learning toolkit

Memberships: