Kevin Sebastian Luck

Assistant Professor

I am an Assistant Professor in the Computatioal Intelligence Group at the Vrije Universiteit Amsterdam. Here, I am working on multiple different topics in the field of robot learning, primarily in the area of deep reinforcement learning.

One area of particular interest of mine is the combination of deep reinforcement learning with evolutionary principles, for the development of AI-assisted design methodologies and physical/embodied intelligence, to allow robots to co-adapt both their behaviour and morpoholgy.

Before I joined Amsterdam I worked as postdoctoral researcher at the Aalto University and the Finnish Center for Artificial Intelligence in Espoo (Helsinki), Finland, where I am/was hosted by the Intelligent Robotics Group of Ville Kyrki.

Keywords: (Deep) Reinforcement Learning, Robot Learning, Embodied Intelligence, Evolving Robots

news

May 05, 2024	I am a recipient of the NWO AIned XS grant provided by the Dutch Research Council and the AIned Foundation. In this projct we will be developing new methods using Large Language Models for policy generation for real-world robotics. Sounds interesting? Keep a look out for our postdoc advertisement and more details in the near future!
May 05, 2024	I will be giving an invited talk about our research on co-adapting robot morphology and behaviour with reinforcement learning at the EARL workshop. Are you also working on Co-Adaptation/Co-Design of robots or evolutionary robots? Consider submitting a workshop paper and let’s chat at R:SS 2024 in Delft, The Netherlands!
Feb 16, 2024	Maria Sliacka’s paper about Co-Imagination, model-based Co-Design of agent’s body and behaviour with dreaming/data-augmentation, is published by Springer. Congratulations!
Feb 01, 2024	I will be Editor for the IEEE International Conference on Intelligent Robotics and Systems (IROS) 2024
Nov 20, 2023	Currently evaluating reinforcement learning on biological systems: I am on parental leave until February 2024
Nov 15, 2023	I am one of three finalists for the Outstanding Associate Editor Award at the IEEE International Conference on Intelligent Robotics and Systems 20203
Sep 20, 2023	Our papers on disentangled representations for RL and Equivariances in Neural Processes got accepted to NeurIPS 2023!
Sep 20, 2023	I presented our paper on Model-Based Co-Adaptation at LOD 2023
Sep 08, 2023	I am a co-organizer of the ELLIS Robust ML Workshop.
Sep 07, 2023	I joined the ELLIS society (European Laboratory for Learning and Intelligent Systems).
Sep 01, 2023	I started my new position as Assistant Professor of Computational Intelligence at the Vrije Universiteit Amsterdam in the Netherlands.
Apr 07, 2023	We are going to present our paper Co-Imitation: Learning Design and Behaviour by Imitation at AAAI-23 in Washington DC. Visit our oral talk in the HAI: Human-in-the-Loop Machine Learning track at 2PM, or visit us at poster 166 on Thursday, 9th of February.
Feb 10, 2023	We are going to present our paper Co-Imitation: Learning Design and Behaviour by Imitation at AAAI-23 in Washington DC. Visit our oral talk in the HAI: Human-in-the-Loop Machine Learning track at 2PM, or visit us at poster 166 on Thursday, 9th of February.
Jan 07, 2023	Our paper about learning disentangled embeddings with Mhairi Dunion Temporal Disentanglement of Representations for Improved Generalisation in Reinforcement Learning was accepted for ICLR 2023!
Dec 12, 2022	I gave an invited talk about our work in the area of co-adaptation of robot morphology and behaviour at the AI Initiave at KAUST, Saudi Arabia.

selected publications

AAAI

Sparse latent space policy search

Kevin Sebastian Luck , Joni Pajarinen , Erik Berger , Ville Kyrki , and Heni Ben Amor

In Proceedings of the AAAI Conference on Artificial Intelligence , 2016

Bib PDF

@inproceedings{luck2016sparse,
  title = {Sparse latent space policy search},
  author = {Luck, Kevin Sebastian and Pajarinen, Joni and Berger, Erik and Kyrki, Ville and Amor, Heni Ben},
  booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
  volume = {30},
  number = {1},
  year = {2016},
}

R:SS

From the Lab to the Desert: Fast Prototyping and Learning of Robot Locomotion

Kevin Sebastian Luck , Joseph Campbell , Michael Jansen , Daniel Aukes , and Heni Ben Amor

In Proceedings of Robotics: Science and Systems , Jul 2017

Bib PDF

@inproceedings{Luck-RSS-17,
  author = {Luck, Kevin Sebastian and Campbell, Joseph and Jansen, Michael and Aukes, Daniel and Amor, Heni Ben},
  title = {From the Lab to the Desert: Fast Prototyping and Learning of Robot Locomotion},
  booktitle = {Proceedings of Robotics: Science and Systems},
  year = {2017},
  address = {Cambridge, Massachusetts},
  month = jul,
  doi = {10.15607/RSS.2017.XIII.075},
}

CoRL
Data-efficient Co-Adaptation of Morphology and Behaviour with Deep Reinforcement Learning

Kevin Sebastian Luck , Heni Ben Amor , and Roberto Calandra

In Proceedings of the Conference on Robot Learning , 30 oct–01 nov 2020

Abs Bib PDF

Humans and animals are capable of quickly learning new behaviours to solve new tasks. Yet, we often forget that they also rely on a highly specialized morphology that co-adapted with motor control throughout thousands of years. Although compelling, the idea of co-adapting morphology and behaviours in robots is often unfeasible because of the long manufacturing times, and the need to redesign an appropriate controller for each morphology. In this paper, we propose a novel approach to automatically and efficiently co-adapt a robot morphology and its controller. Our approach is based on recent advances in deep reinforcement learning, and specifically the soft actor critic algorithm. Key to our approach is the possibility of leveraging previously tested morphologies and behaviors to estimate the performance of new candidate morphologies. As such, we can make full use of the information available for making more informed decisions, with the ultimate goal of achieving a more data-efficient co-adaptation (i.e., reducing the number of morphologies and behaviors tested). Simulated experiments show that our approach requires drastically less design prototypes to find good morphology-behaviour combinations, making this method particularly suitable for future co-adaptation of robot designs in the real world.
@inproceedings{pmlr-v100-luck20a, title = {Data-efficient Co-Adaptation of Morphology and Behaviour with Deep Reinforcement Learning}, author = {Luck, Kevin Sebastian and Amor, Heni Ben and Calandra, Roberto}, booktitle = {Proceedings of the Conference on Robot Learning}, pages = {854--869}, year = {2020}, editor = {Kaelbling, Leslie Pack and Kragic, Danica and Sugiura, Komei}, volume = {100}, series = {Proceedings of Machine Learning Research}, month = {30 Oct--01 Nov}, publisher = {PMLR}, url = {https://proceedings.mlr.press/v100/luck20a.html}, }

NeurIPS

Conditional Mutual Information for Disentangled Representations in Reinforcement Learning

Mhairi Dunion , Trevor McInroe , Kevin Sebastian Luck , Josiah P. Hanna , and Stefano V Albrecht

In Thirty-seventh Conference on Neural Information Processing Systems , 30 oct–01 nov 2023

Bib PDF

@inproceedings{dunion2023conditional,
  title = {Conditional Mutual Information for Disentangled Representations in Reinforcement Learning},
  author = {Dunion, Mhairi and McInroe, Trevor and Luck, Kevin Sebastian and Hanna, Josiah P. and Albrecht, Stefano V},
  booktitle = {Thirty-seventh Conference on Neural Information Processing Systems},
  year = {2023},
  url = {https://openreview.net/forum?id=EmYWJsyad4},
}

AAAI

Co-imitation: learning design and behaviour by imitation

Chang Rajani , Karol Arndt , David Blanco-Mulero , Kevin Sebastian Luck , and Ville Kyrki

In Proceedings of the AAAI Conference on Artificial Intelligence , 30 oct–01 nov 2023

Bib PDF

@inproceedings{rajani2023co,
  title = {Co-imitation: learning design and behaviour by imitation},
  author = {Rajani, Chang and Arndt, Karol and Blanco-Mulero, David and Luck, Kevin Sebastian and Kyrki, Ville},
  booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
  volume = {37},
  number = {5},
  pages = {6200--6208},
  year = {2023},
}