Luca Grillotti

Holà! I’m Luca Grillotti, a PhD student at the Adaptive and Intelligent Robotics Lab (AIRL) of Imperial College London under the supervision of Antoine Cully.

My research focuses on how to use Reinforcement Learning and Quality-Diversity algorithms to learn collections of diverse and high-performing robot abilities in an unsupervised manner.

I also work as Teaching Scholar for the Department of Computing. I had the opportunity to give several PyTorch lectures and design the corresponding lecture notes, teach half of a Computational Techniques module, create many Python Programming and Probabilistic Inference assignments from scratch, and work as teaching assistant for the Reinforcement Learning and Mathematics modules.

While working as a Teaching Scholar, I obtained a Postgraduate Certificate in University Learning and Teaching (PGCert ULT) in 2022. The PGCert programme included several courses about teaching and learning in the university setting, such as: approaches to teaching, how students learn, digital learning and educational supervision.

This poster summarises the research conducted during my first months of PhD.

Selected Publications

Unsupervised Behavior Discovery With Quality-Diversity Optimization

Luca Grillotti, and Antoine Cully

IEEE Transactions on Evolutionary Computation, 2022

Bib HTML PDF

@article{grillotti2022unsupervised,
  title = {Unsupervised Behavior Discovery With Quality-Diversity Optimization},
  author = {Grillotti, Luca and Cully, Antoine},
  journal = {IEEE Transactions on Evolutionary Computation},
  volume = {26},
  number = {6},
  pages = {1539--1552},
  year = {2022},
  publisher = {IEEE},
}

Don’t Bet on Luck Alone: Enhancing Behavioral Reproducibility of Quality-Diversity Solutions in Uncertain Domains

Luca Grillotti*, Manon Flageat*, Bryan Lim, and 1 more author

In Proceedings of the Genetic and Evolutionary Computation Conference, 2023

PDF

Relevance-guided unsupervised discovery of abilities with quality-diversity algorithms

Luca Grillotti, and Antoine Cully

In Proceedings of the Genetic and Evolutionary Computation Conference, 2022

Bib HTML PDF

@inproceedings{grillotti2022relevance,
  title = {Relevance-guided unsupervised discovery of abilities with quality-diversity algorithms},
  author = {Grillotti, Luca and Cully, Antoine},
  booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference},
  pages = {77--85},
  year = {2022},
}

ICLR
ALOE Workshop

Discovering Unsupervised Behaviours from Full State Trajectories

Luca Grillotti, and Antoine Cully

In ICLR Workshop on Agent Learning in Open-Endedness, 2022

Bib PDF

@inproceedings{grillotti2022discovering,
  title = {Discovering Unsupervised Behaviours from Full State Trajectories},
  author = {Grillotti, Luca and Cully, Antoine},
  booktitle = {ICLR Workshop on Agent Learning in Open-Endedness},
  year = {2022},
}