About - Frederik Schubert

My name is Frederik Schubert If you want to get in touch with me you can send me an email to info@frederik-schubert.de. and I’m a PhD student at the Institute for Information Processing at Leibniz University and the Head of Data Science at INSIDE M2M in Hannover, Germany.

As a software engineer, my experience ranges from the backend to the frontend, as well as devops and infrastructure, especially in a machine learning context. My research focuses on data-efficient methods for reinforcement learning, computer vision and, recently, inverse design for photonic integrated circuits.

Publications

@misc{reinders2025hydramixmultiimagefeaturemixing,
      title={HydraMix: Multi-Image Feature Mixing for Small Data Image Classification}, 
      author={Christoph Reinders and Frederik Schubert and Bodo Rosenhahn},
      year={2025},
      eprint={2501.09504},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2501.09504}, 
}

@article{schubert2025quantized,
  author  = {Schubert, Frederik and Mahlau, Yannik and Bethmann, Konrad and Hartmann, Fabian and Caspary, Reinhard and Munderloh, Marco and Ostermann, J{\"o}rn and Rosenhahn, Bodo},
  title   = {Quantized Inverse Design for Photonic Integrated Circuits},
  journal = {ACS Omega},
  year    = {2025},
  doi     = {10.1021/acsomega.4c10958},
  url     = {https://doi.org/10.1021/acsomega.4c10958},
  note    = {ACS Omega}
}

@misc{mahlau2024flexibleframeworklargescalefdtd,
      title={A flexible framework for large-scale FDTD simulations: open-source inverse design for 3D nanostructures}, 
      author={Yannik Mahlau and Frederik Schubert and Konrad Bethmann and Reinhard Caspary and Antonio Calà Lesina and Marco Munderloh and Jörn Ostermann and Bodo Rosenhahn},
      year={2024},
      eprint={2412.12360},
      archivePrefix={arXiv},
      primaryClass={physics.optics},
      url={https://arxiv.org/abs/2412.12360}, 
      note = {Photonics West 2025}
}

@inproceedings{mahlau2024mastering,
author = {Mahlau, Yannik and Schubert, Frederik and Rosenhahn, Bodo},
title = {Mastering zero-shot interactions in cooperative and competitive simultaneous games},
year = {2024},
publisher = {JMLR.org},
abstract = {The combination of self-play and planning has achieved great successes in sequential games, for instance in Chess and Go. However, adapting algorithms such as AlphaZero to simultaneous games poses a new challenge. In these games, missing information about concurrent actions of other agents is a limiting factor as they may select different Nash equilibria or do not play optimally at all. Thus, it is vital to model the behavior of the other agents when interacting with them in simultaneous games. To this end, we propose Albatross: AlphaZero for Learning Bounded-rational Agents and Temperature-based Response Optimization using Simulated Self-play. Albatross learns to play the novel equilibrium concept of a Smooth Best Response Logit Equilibrium (SBRLE), which enables cooperation and competition with agents of any playing strength. We perform an extensive evaluation of Albatross on a set of cooperative and competitive simultaneous perfect-information games. In contrast to AlphaZero, Albatross is able to exploit weak agents in the competitive game of Battlesnake. Additionally, it yields an improvement of 37.6\% compared to previous state of the art in the cooperative Overcooked benchmark.},
booktitle = {Proceedings of the 41st International Conference on Machine Learning},
articleno = {1394},
numpages = {28},
location = {Vienna, Austria},
series = {ICML'24},
note   = {ICML 2024},
url = {https://dl.acm.org/doi/abs/10.5555/3692070.3693464}
}

@article{benjamins2022ContextualizeM,
  title = {Contextualize Me \textendash{} the Case for Context in Reinforcement Learning},
  author = {Benjamins, Carolin and Eimer, Theresa and Schubert, Frederik and Mohan, Aditya and D{\"o}hler, Sebastian and Biedenkapp, Andr{\'e} and Rosenhahn, Bodo and Hutter, Frank and Lindauer, Marius},
  year = {2023},
  journal = {Transactions on Machine Learning Research},
  note = {Transactions on Machine Learning Research},
  issn = {2835-8856},
  url     = {https://openreview.net/forum?id=Y42xVBQusn}
}

@article {SchBen2022,
  title = {{{POLTER}}: {{Policy}} Trajectory Ensemble Regularization for Unsupervised Reinforcement Learning},
  author = {Schubert, Frederik and Benjamins, Carolin and D{\"o}hler, Sebastian and Rosenhahn, Bodo and Lindauer, Marius},
  year = {2023},
  journal = {Transactions on Machine Learning Research},
  note = {Transactions on Machine Learning Research},
  issn = {2835-8856},
  url = {https://openreview.net/forum?id=Hnr23knZfY},
}

@article{awiszus2022WorldGANTN,
  title   = {Wor(l)d-GAN: Towards Natural Language Based PCG in Minecraft},
  author  = {Maren Awiszus and Frederik Schubert and Bodo Rosenhahn},
  journal = {IEEE Transactions on Games},
  year    = {2022},
  url     = {https://ieeexplore.ieee.org/document/9720203},
  note    = {IEEE Transactions on Games}
}

@inproceedings{reinders2022ImageCO,
  title  = {Image Classification on Small Datasets via Masked Feature Mixing},
  author = {Christoph Reinders and Frederik Schubert and Bodo Rosenhahn},
  year   = {2022},
  booktitle = {IJCAI-ECAI 2021},
  url    = {https://arxiv.org/abs/2202.11616},
  note   = {IJCAI-ECAI 2022}
}

@article{benjaminsCARLBenchmarkContextual2021,
  title         = {{{CARL}}: A {{Benchmark}} for {{Contextual}} and {{Adaptive Reinforcement Learning}}},
  shorttitle    = {{{CARL}}},
  author        = {Benjamins, Carolin and Eimer, Theresa and Schubert, Frederik and Biedenkapp, Andr{\'e} and Rosenhahn, Bodo and Hutter, Frank and Lindauer, Marius},
  year          = {2021},
  month         = oct,
  journal       = {arXiv:2110.02102 [cs]},
  eprint        = {2110.02102},
  eprinttype    = {arxiv},
  primaryclass  = {cs},
  abstract      = {While Reinforcement Learning has made great strides towards solving ever more complicated tasks, many algorithms are still brittle to even slight changes in their environment. This is a limiting factor for real-world applications of RL. Although the research community continuously aims at improving both robustness and generalization of RL algorithms, unfortunately it still lacks an open-source set of well-defined benchmark problems based on a consistent theoretical framework, which allows comparing different approaches in a fair, reliable and reproducible way. To fill this gap, we propose CARL, a collection of well-known RL environments extended to contextual RL problems to study generalization. We show the urgent need of such benchmarks by demonstrating that even simple toy environments become challenging for commonly used approaches if different contextual instances of this task have to be considered. Furthermore, CARL allows us to provide first evidence that disentangling representation learning of the states from the policy learning with the context facilitates better generalization. By providing variations of diverse benchmarks from classic control, physical simulations, games and a real-world application of RNA design, CARL will allow the community to derive many more such insights on a solid empirical foundation.},
  archiveprefix = {arXiv},
  langid        = {english},
  annotation    = {ZSCC: 0000000},
  note          = {Ecological Theory of RL Workshop @ NeurIPS 2021},
  url           = {https://arxiv.org/abs/2110.02102}
}

@inproceedings{awiszus2021world,
  title     = {World-GAN: a Generative Model for Minecraft Worlds},
  author    = {Awiszus, Maren and Schubert, Frederik and Rosenhahn, Bodo},
  booktitle = {2021 IEEE Conference on Games (CoG)},
  year      = {2021},
  url       = {https://arxiv.org/abs/2106.10155},
  note      = {IEEE Conference on Games 2021}
}

@article{schubertAutomaticRiskAdaptation2021,
  title         = {Automatic {{Risk Adaptation}} in {{Distributional Reinforcement Learning}}},
  author        = {Schubert, Frederik and Eimer, Theresa and Rosenhahn, Bodo and Lindauer, Marius},
  year          = {2021},
  month         = jun,
  journal       = {arXiv:2106.06317 [cs]},
  eprint        = {2106.06317},
  eprinttype    = {arxiv},
  primaryclass  = {cs},
  abstract      = {The use of Reinforcement Learning (RL) agents in practical applications requires the consideration of suboptimal outcomes, depending on the familiarity of the agent with its environment. This is especially important in safety-critical environments, where errors can lead to high costs or damage. In distributional RL, the risksensitivity can be controlled via different distortion measures of the estimated return distribution. However, these distortion functions require an estimate of the risk level, which is difficult to obtain and depends on the current state. In this work, we demonstrate the suboptimality of a static risk level estimation and propose a method to dynamically select risk levels at each environment step. Our method ARA (Automatic Risk Adaptation) estimates the appropriate risk level in both known and unknown environments using a Random Network Distillation error. We show reduced failure rates by up to a factor of 7 and improved generalization performance by up to 14\% compared to both risk-aware and risk-agnostic agents in several locomotion environments.},
  archiveprefix = {arXiv},
  langid        = {english},
  annotation    = {ZSCC: 0000000},
  note          = {Reinforcement Learning for Real Life Workshop @ ICML 2021},
  url           = {https://arxiv.org/abs/2106.06317}
}

@article{schubert2021toad,
  author  = {Schubert, Frederik and Awiszus, Maren and Rosenhahn, Bodo},
  journal = {IEEE Transactions on Games},
  title   = {TOAD-GAN: a Flexible Framework for Few-Shot Level Generation in Token-Based Games},
  year    = {2021},
  volume  = {},
  number  = {},
  pages   = {1-1},
  doi     = {10.1109/TG.2021.3069833},
  url     = {https://ieeexplore.ieee.org/document/9390320},
  note    = {IEEE Transactions on Games}
}

@inproceedings{awiszus2020toad,
  title     = {TOAD-GAN: coherent style level generation from a single example},
  author    = {Awiszus, Maren and Schubert, Frederik and Rosenhahn, Bodo},
  booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment},
  volume    = {16},
  number    = {1},
  pages     = {10--16},
  year      = {2020},
  note      = {AIIDE 2020, Best Student Paper Award},
  url       = {https://ojs.aaai.org/index.php/AIIDE/article/download/7401/7332}
}