Sayantan Auddy

Hi! I’m Sayantan, a postdoctoral researcher at the Learning and Intelligent Systems (LIS) Lab led by Prof. Dr. Marc Toussaint at Technische Universität Berlin, Germany. My current research focuses on reinforcement learning and sim-to-real transfer in robotics, including developing mechanisms for effective zero-shot transfer and fast, data-efficient adaptation on real systems to handle changes after deployment.

I earned my PhD from Universität Innsbruck, Austria, under the supervision of Prof. Dr. Justus Piater, where I worked on continual learning for manipulation, learning from demonstration, and sim-to-real transfer. Previously, I completed my MSc (with distinction) at Universität Hamburg, Germany, working closely with Prof. Dr. Stefan Wermter.

If our interests align, feel free to reach out — I’d love to connect!

Selected Research

(Please visit my Google Scholar profile for the full publication list.)

CrazyMARL: Decentralized Direct Motor Control Policies for Cooperative Aerial Transport of Cable-Suspended Payloads
V. Lorentz, K. Wahba, S. Auddy, M. Toussaint, W. Hönig
Preprint, under review, 2025.
Paper |

BibTeX

@article{crazymarl2025,
  title={CrazyMARL: Decentralized Direct Motor Control Policies for Cooperative Aerial Transport of Cable-Suspended Payloads},
  author={Lorentz, V. and Wahba, K. and Auddy, S. and Toussaint, M. and Hönig, W.},
  year={2025},
  note={Preprint}
}

Trajectory First: A Curriculum for Discovering Diverse Policies
C.V. Braun, S. Auddy, M. Toussaint
Inductive Biases in Reinforcement Learning Workshop at RLC, 2025. Best Paper Award.
Paper |

BibTeX

@inproceedings{trajectory2025,
  title={Trajectory First: A Curriculum for Discovering Diverse Policies},
  author={Braun, C.V. and Auddy, S. and Toussaint, M.},
  booktitle={Inductive Biases in Reinforcement Learning Workshop at RLC},
  year={2025}
}

Scalable and Efficient Continual Learning from Demonstration via Hypernetwork-generated Stable Dynamics Model
S. Auddy, J. Hollenstein, M. Saveriano, A. Rodríguez-Sánchez, J. Piater
Preprint, under review, 2025.
Paper |

BibTeX

@article{hypernet2025,
  title={Scalable and Efficient Continual Learning from Demonstration via Hypernetwork-generated Stable Dynamics Model},
  author={Auddy, S. and Hollenstein, J. and Saveriano, M. and Rodríguez-Sánchez, A. and Piater, J.},
  year={2025},
  note={Preprint}
}

Imitation learning-based Direct Visual Servoing using the large projection formulation
S. Auddy, A. Paolillo, J. Piater, M. Saveriano
Robotics and Autonomous Systems (RAS), 2025.
Paper |

BibTeX

@article{visualservo2025,
  title={Imitation learning-based Direct Visual Servoing using the large projection formulation},
  author={Auddy, S. and Paolillo, A. and Piater, J. and Saveriano, M.},
  journal={Robotics and Autonomous Systems},
  year={2025}
}

Safe Continual Domain Adaptation after Sim2Real Transfer of Reinforcement Learning Policies in Robotics
J. Josifovski, S. Gu, M. Malmir, H. Huang, S. Auddy, N. Navarro-Guerrero, C. Spanos, A. Knoll
Preprint, 2025.
Paper |

BibTeX

@article{safecda2025,
  title={Safe Continual Domain Adaptation after Sim2Real Transfer of Reinforcement Learning Policies in Robotics},
  author={Josifovski, J. and Gu, S. and Malmir, M. and Huang, H. and Auddy, S. and Navarro-Guerrero, N. and Spanos, C. and Knoll, A.},
  year={2025},
  note={Preprint}
}

Continual Domain Randomization
J. Josifovski*, S. Auddy*, M. Malmir, J. Piater, A. Knoll, N. Navarro-Guerrero
International Conference on Intelligent Robots and Systems (IROS), 2024.
(* denotes equal contribution)
Paper |

BibTeX

@inproceedings{domainrand2024,
  title={Continual Domain Randomization},
  author={Josifovski, J. and Auddy, S. and Malmir, M. and Piater, J. and Knoll, A. and Navarro-Guerrero, N.},
  booktitle={IROS},
  year={2024}
}

Continual Learning from Demonstration of Robotics Skills
S. Auddy, J. Hollenstein, M. Saveriano, A. Rodríguez-Sánchez, J. Piater
Robotics and Autonomous Systems (RAS), Conference on Lifelong Learning Agents (CoLLAs), 2023.
Paper |

BibTeX

@article{cld2023,
  title={Continual Learning from Demonstration of Robotics Skills},
  author={Auddy, S. and Hollenstein, J. and Saveriano, M. and Rodríguez-Sánchez, A. and Piater, J.},
  journal={Robotics and Autonomous Systems},
  year={2023}
}

Action Noise in Off-Policy Deep Reinforcement Learning: Impact on Exploration and Performance
J. Hollenstein, S. Auddy, M. Saveriano, E. Renaudo, J. Piater
Transactions on Machine Learning Research (TMLR), 2022.
Paper |

BibTeX

@article{actionnoise2022,
  title={Action Noise in Off-Policy Deep Reinforcement Learning: Impact on Exploration and Performance},
  author={Hollenstein, J. and Auddy, S. and Saveriano, M. and Renaudo, E. and Piater, J.},
  journal={Transactions on Machine Learning Research},
  year={2022}
}

Hierarchical control for bipedal locomotion using central pattern generators and neural networks
S. Auddy, S. Magg, S. Wermter
International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob), 2018. Oral.
Paper |

BibTeX

@inproceedings{hierarchical2018,
  title={Hierarchical control for bipedal locomotion using central pattern generators and neural networks},
  author={Auddy, S. and Magg, S. and Wermter, S.},
  booktitle={ICDL-EpiRob},
  year={2018}
}