Leander Diaz-Bone

I am a PhD student in the Center for Learning Systems program advised by Andreas Krause at ETH Zürich and Bernhard Schölkopf at the MPI for Intelligent Systems in Tübingen. Prior to this, I completed my Bachelor’s and Master’s degrees at ETH Zürich, which included an exchange quarter at the University of Washington.

My research focuses on scalable and sample-efficient learning from interaction. I am currently working on world models that learn abstract representations of their environment.

Feel free to reach out to me about research interests and ideas; I am always happy to chat. (ldiazbone@ethz.ch)

news

Sep 20, 2025	I am happy to share that DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning has been accepted at NeurIPS 2025.

selected publications

Learning on the Job: Test-Time Curricula for Targeted Reinforcement Learning

Jonas Hübotter^*, Leander Diaz-Bone^*, Ido Hakimi, and 2 more authors

2025

arXiv Bib PDF Code

@misc{huebotter2025learningjobtesttimecurricula,
  title = {Learning on the Job: Test-Time Curricula for Targeted Reinforcement Learning},
  author = {H{\"u}botter, Jonas and Diaz-Bone, Leander and Hakimi, Ido and Krause, Andreas and Hardt, Moritz},
  year = {2025},
  eprint = {2510.04786},
  archiveprefix = {arXiv},
  primaryclass = {cs.LG},
}

DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning

Leander Diaz-Bone^*, Marco Bagatella^*, Jonas Hübotter^*, and 1 more author

In Advances in Neural Information Processing Systems (NeurIPS), Dec 2025

arXiv Bib PDF Code

@inproceedings{diazbone2025discover,
  author = {Diaz-Bone, Leander and Bagatella, Marco and H{\"u}botter, Jonas and Krause, Andreas},
  title = {DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  month = dec,
  year = {2025},
}