Dr. Carsten Hahn

Dr. Carsten Hahn

Lehrstuhl für Mobile und Verteilte Systeme

Ludwig-Maximilians-Universität München, Institut für Informatik

Oettingenstraße 67
80538 München

Raum E 008

Telefon: +49 89 / 2180-9152

Fax: +49 89 / 2180-9148

Mail: carsten.hahn@ifi.lmu.de

Forschungsinteressen

  • Multi-Agent Systems
  • Reinforcement Learning
  • Artificial Neural Networks
  • Intuitive Physics
  • Route Planning and Navigation
  • Anomaly Detection

Abschlussarbeiten

 

Bei Interesse an einer Abschlussarbeit in meinen Themengebieten oder auch bei eigenen Ideen könnt Ihr euch gerne bei mir melden. Die Liste der aktuell ausgeschriebenen Abschlussarbeiten ist nicht zwingend immer vollständig.

Generell sind Vorkenntnisse in Python, künstlichen neuronalen Netzen (in Keras bzw. TensorFlow), Reinforcement Learning und Graphalgorithmen bei der Bearbeitung meiner Themen von Vorteil.

 

Ausgeschriebene Themen

–

Abgeschlossene Arbeiten

  • Ein vorausschauender Kollisionsmanagementalgorithmus für eine zeit- und risikoabhängige Pfadplanung
  • Steuerung multipler Agenten in einer Räuber-Beute-Umgebung mittels Deep Reinforcement Learning auf Bilddaten
  • Steuerung eines Schwarms autonomer Agenten mithilfe von Reinforcement Learning
  • Koevolution in einer kontinuierlichen Räuber-Beute-Umgebung mittels Multi-Agent Reinforcement Learning
  • Nearest Neighbor Search mit künstlichen neuronalen Netzen
  • Intuitive Vorhersage der Bewegung dynamischer Objekte mit neuronalen Netzen
  • Neuronales Gas für dynamische Pfadplanung
  • Ein Vergleich verschiedener Architekturen künstlicher neuronaler Netze zur Anomalieerkennung in Zeitserien
  • Problems with Permutation Invariance in Neural Networks
  • Abschätzung physikalischer Gegebenheiten mit neuronalen Netzen unter Unsicherheit
  • Einbettung ausfallsicherer Netzwerkservices

Lehre

  • Betriebssysteme (WS 2015/2016, WS 2016/2017, WS 2017/2018, WS 2018/2019, WS 2019/2020, WS 2020/2021)
  • Rechnerarchitektur (SS 2016, SS 2017, SS 2018, SS 2019, SS 2020)
  • Praktikum iOS-Entwicklung (WS 2016/2017)

Publikationen

2020

  • C. Hahn, F. Ritz, P. Wikidal, T. Phan, T. Gabor, and C. Linnhoff-Popien, „Foraging Swarms using Multi-Agent Reinforcement Learning,“ in Conference on Artificial Life (ALIFE 2020), 2020.
    [BibTeX] 
    @inproceedings{hahn2020swarmcontrol,
author = {Carsten Hahn and Fabian Ritz and Paula Wikidal and Thomy Phan and Thomas Gabor and Claudia Linnhoff-Popien},
title = {Foraging Swarms using Multi-Agent Reinforcement Learning},
booktitle = {Conference on Artificial Life (ALIFE 2020)},
year = {2020},
owner = {chahn}
}

  • C. Roch, T. Phan, S. Feld, R. Müller, T. Gabor, C. Hahn, and C. Linnhoff-Popien, „A Quantum Annealing Algorithm for Finding Pure Nash Equilibria in Graphical Games,“ in 20th International Conference on Computational Science (ICCS 2020), 2020, p. 12. doi:10.1007/978-3-030-50433-5_38
    [BibTeX] [Download PDF] 
    @inproceedings{roch2020quantum,
author = {Christoph Roch and Thomy Phan and Sebastian Feld and Robert Müller and Thomas Gabor and Carsten Hahn and Claudia Linnhoff-Popien},
title = {A Quantum Annealing Algorithm for Finding Pure Nash Equilibria in Graphical Games},
booktitle = {20th International Conference on Computational Science (ICCS 2020)},
year = {2020},
month = {June},
pages = {12},
url = {https://www.iccs-meeting.org/archive/iccs2020/papers/121420466.pdf},
doi = {10.1007/978-3-030-50433-5_38}
}

  • F. Ritz, F. Hohnstein, R. Müller, T. Phan, T. Gabor, C. Hahn, and C. Linnhoff-Popien, „Towards Ecosystem Management from Greedy Reinforcement Learning in a Predator-Prey Setting,“ in Artificial Life Conference Proceedings, 2020, pp. 518-525. doi:10.1162/isal_a_00273
    [BibTeX] [Abstract] [Download PDF]

    This paper applies reinforcement learning to train a predator to hunt multiple prey, which are able to reproduce, in a 2D simulation. It is shown that, using methods of curriculum learning, long-term reward discounting and stacked observations, a reinforcement-learning-based predator can achieve an economic strategy: Only hunt when there is still prey left to reproduce in order to maintain the population. Hence, purely selfish goals are sufficient to motivate a reinforcement learning agent for long-term planning and keeping a certain balance with its environment by not depleting its resources. While a comparably simple reinforcement learning algorithm achieves such behavior in the present scenario, providing a suitable amount of past and predictive information turns out to be crucial for the training success.

    @inproceedings{ritz20towards,
author = {Ritz, Fabian and Hohnstein, Felix and Müller, Robert and Phan, Thomy and Gabor, Thomas and Hahn, Carsten and Linnhoff-Popien, Claudia},
title = {Towards Ecosystem Management from Greedy Reinforcement Learning in a Predator-Prey Setting},
booktitle = {Artificial Life Conference Proceedings},
volume = {32},
pages = {518-525},
year = {2020},
doi = {10.1162/isal\_a\_00273},
url = {https://www.mitpressjournals.org/doi/abs/10.1162/isal_a_00273},
abstract = {This paper applies reinforcement learning to train a predator to hunt multiple prey, which are able to reproduce, in a 2D simulation. It is shown that, using methods of curriculum learning, long-term reward discounting and stacked observations, a reinforcement-learning-based predator can achieve an economic strategy: Only hunt when there is still prey left to reproduce in order to maintain the population. Hence, purely selfish goals are sufficient to motivate a reinforcement learning agent for long-term planning and keeping a certain balance with its environment by not depleting its resources. While a comparably simple reinforcement learning algorithm achieves such behavior in the present scenario, providing a suitable amount of past and predictive information turns out to be crucial for the training success.}
}

  • M. Friedrich, C. Roch, S. Feld, C. Hahn, and P. Fayolle, „A Flexible Pipeline for the Optimization of Construction Trees,“ in Proceedings of the 28th International Conference on Computer Graphics, Visualization and Computer Vision (WSCG), 2020.
    [BibTeX] 
    @inproceedings{friedrich2020csg-optim,
author = {Friedrich, Markus and Roch, Christoph and Feld, Sebastian and Hahn, Carsten and Fayolle, Pierre-Alain},
title = {A Flexible Pipeline for the Optimization of Construction Trees},
booktitle = {Proceedings of the 28th International Conference on Computer Graphics, Visualization and Computer Vision (WSCG)},
year = {2020},
owner = {mfriedrich}
}

  • C. Hahn, T. Phan, S. Feld, C. Roch, F. Ritz, A. Sedlmeier, T. Gabor, and C. Linnhoff-Popien, „Nash Equilibria in Multi-Agent Swarms,“ in 12th International Conference on Agents and Artificial Intelligence (ICAART 2020), 2020.
    [BibTeX] 
    @inproceedings{hahn2020nash,
author = {Carsten Hahn and Thomy Phan and Sebastian Feld and Christoph Roch and Fabian Ritz and Andreas Sedlmeier and Thomas Gabor and Claudia Linnhoff-Popien},
title = {Nash Equilibria in Multi-Agent Swarms},
booktitle = {12th International Conference on Agents and Artificial Intelligence (ICAART 2020)},
year = {2020},
owner = {chahn}
}

  • C. Hahn, S. Feld, and H. Schroter, „Predictive Collision Management for Time and Risk Dependent Path Planning,“ in 28th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (SIGSPATIAL 2020), 2020.
    [BibTeX] 
    @inproceedings{hahn2020sigspatial,
author = {Hahn, Carsten and Feld, Sebastian and Schroter, Hannes},
title = {Predictive Collision Management for Time and Risk Dependent Path Planning},
booktitle = {28th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems (SIGSPATIAL 2020)},
year = {2020},
organization = {ACM},
owner = {chahn}
}

2019

  • C. Hahn, S. Feld, M. Zierl, and C. Linnhoff-Popien, „Dynamic Path Planning with Stable Growing Neural Gas,“ in 11th International Conference on Agents and Artificial Intelligence (ICAART 2019), 2019.
    [BibTeX] 
    @inproceedings{hahn2019dynamic,
author = {Carsten Hahn and Sebastian Feld and Manuel Zierl and Claudia Linnhoff-Popien},
title = {Dynamic Path Planning with Stable Growing Neural Gas},
booktitle = {11th International Conference on Agents and Artificial Intelligence (ICAART 2019)},
year = {2019},
owner = {chahn}
}

  • C. Hahn, T. Phan, T. Gabor, L. Belzner, and C. Linnhoff-Popien, „Emergent Escape-based Flocking Behavior using Multi-Agent Reinforcement Learning,“ in Conference on Artificial Life (ALIFE 2019), 2019.
    [BibTeX] 
    @inproceedings{hahn2019swarm,
author = {Carsten Hahn and Thomy Phan and Thomas Gabor and Lenz Belzner and Claudia Linnhoff-Popien},
title = {Emergent Escape-based Flocking Behavior using Multi-Agent Reinforcement Learning},
booktitle = {Conference on Artificial Life (ALIFE 2019)},
year = {2019},
owner = {chahn}
}

  • M. Friedrich, A. Ebert, C. Hahn, G. Schneider, L. Obermeier, A. Erk, and I. Jennes, „A Distributed Metadata Platform for Hybrid Radio Services,“ in 19th International Conference on Innovations for Community Services (I4CS 2019), 2019.
    [BibTeX] 
    @inproceedings{friedrich2019distributed-metadata,
title = {A Distributed Metadata Platform for Hybrid Radio Services},
author = {Friedrich, Markus and Ebert, André and Hahn, Carsten and Schneider, Georg and Obermeier, Liza and Erk, Alexander and Jennes, Iris},
booktitle = {19th International Conference on Innovations for Community Services (I4CS 2019)},
year = {2019},
owner = {mfriedrich}
}

  • C. Hahn and M. Friedrich, „Using Existing Reinforcement Learning Libraries in Multi-Agent Scenarios,“ in 1st International Symposium on Applied Artificial Intelligence (ISAAI’19), 2019.
    [BibTeX] 
    @inproceedings{hahn2019existing,
author = {Carsten Hahn and Markus Friedrich},
title = {Using Existing Reinforcement Learning Libraries in Multi-Agent Scenarios},
booktitle = {1st International Symposium on Applied Artificial Intelligence (ISAAI'19)},
year = {2019},
owner = {chahn}
}

2018

  • C. Hahn and S. Feld, „Collision Avoidance using Intuitive Physics,“ in 2018 International Symposium on Innovations in Intelligent Systems and Applications (INISTA), 2018.
    [BibTeX] 
    @inproceedings{hahn2018collision,
author = {Hahn, Carsten and Feld, Sebastian},
title = {Collision Avoidance using Intuitive Physics},
booktitle = {2018 International Symposium on Innovations in Intelligent Systems and Applications (INISTA)},
year = {2018},
organization = {IEEE},
owner = {chahn}
}

2017

  • C. Hahn, S. Holzner, L. Belzner, and M. T. Beck, „Empirical Evaluation of a Distributed Deployment Strategy for Virtual Networks,“ in International Conference on Mobile, Secure, and Programmable Networking, 2017, p. 88–98.
    [BibTeX] 
    @inproceedings{hahn2017empirical,
author = {Hahn, Carsten and Holzner, Stephan and Belzner, Lenz and Beck, Michael Till},
title = {Empirical Evaluation of a Distributed Deployment Strategy for Virtual Networks},
booktitle = {International Conference on Mobile, Secure, and Programmable Networking},
year = {2017},
pages = {88--98},
publisher = {Springer},
owner = {chahn}
}

  • C. Hahn, „Wegfindung von Agenten mittels Methoden des maschinellen Lernens,“ in 14. GI/ITG KuVS Fachgespräch Ortsbezogene Anwendungen und Dienste (LBAS 2017), 2017.
    [BibTeX] 
    @inproceedings{hahn2017wegfindung,
author = {Hahn, Carsten},
title = {Wegfindung von Agenten mittels Methoden des maschinellen Lernens},
booktitle = {14. GI/ITG KuVS Fachgespräch Ortsbezogene Anwendungen und Dienste (LBAS 2017)},
year = {2017}
}

2016

  • M. Werner, C. Hahn, and L. Schauer, „DeepMoVIPS: Visual Indoor Positioning Using Transfer Learning,“ in 7th International Conference on Indoor Positioning and Indoor Navigation (IPIN 2016), 2016.
    [BibTeX] 
    @inproceedings{werner2016deepmovips,
title = {DeepMoVIPS: Visual Indoor Positioning Using Transfer Learning},
author = {Martin Werner and Carsten Hahn and Lorenz Schauer},
booktitle = {7th International Conference on Indoor Positioning and Indoor Navigation (IPIN 2016)},
year = {2016},
owner = {mwerner}
}

  • C. Hahn, „Framework for Hybrid Positioning,“ in 13. GI/ITG KuVS Fachgespräch Ortsbezogene Anwendungen und Dienste (LBAS 2016), 2016.
    [BibTeX] 
    @inproceedings{hahn2016framework,
author = {Hahn, Carsten},
title = {Framework for Hybrid Positioning},
booktitle = {13. GI/ITG KuVS Fachgespräch Ortsbezogene Anwendungen und Dienste (LBAS 2016)},
year = {2016},
owner = {chahn}
}