Jonas Hübotter

I am a PhD student in the Learning and Adaptive Systems Group at ETH Zurich working with Andreas Krause. Prior to this, I obtained a Master’s degree in Theoretical Computer Science and Machine Learning from ETH Zurich and a Bachelor’s degree in Computer Science and Mathematics from the Technical University of Munich. I am a recipient of the ETH Medal.

My research aims to leverage foundation models for solving hard tasks through specialization and reinforcement learning. Beyond this, I have broad interests including (approximate) probabilistic inference, optimization, and online learning.

Always feel free to reach out to me with things you find exciting.

Contacts: jhuebotter@ethz.ch Google Scholar GitHub Linkedin

Announcements

Jul, 2025	COLM 2025: Local Mixtures of Experts: Essentially Free Test-Time Training via Model Merging has been accepted! We will also present our work on test-time scaling via prefix-confidence at the SCALR workshop.
May, 2025	ICML 2025: Active Fine-Tuning of Multi-Task Policies has been accepted!
Feb, 2025	Very excited to share notes on Probabilistic AI that I have been writing with Andreas Krause!
Jan, 2025	ICLR 2025: Efficiently Learning at Test-Time: Active Fine-Tuning of LLMs has been accepted!
Jan, 2025	AISTATS 2025: LITE: Efficiently Estimating Gaussian Probability of Maximality has been accepted!

Selected Publications

DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning

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

arXiv preprint arXiv:2505.19850, 2025

Bib PDF Code

@article{diazbone2025discover,
  title = {DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning},
  author = {Diaz-Bone, Leander and Bagatella, Marco and Hübotter, Jonas and Krause, Andreas},
  year = {2025},
  journal = {arXiv preprint arXiv:2505.19850},
}

COLM ’25

Local Mixtures of Experts: Essentially Free Test-Time Training via Model Merging

Ryo Bertolissi^*, Jonas Hübotter^*, Ido Hakimi , and 1 more author

In Conference on Language Modeling , 2025

Bib PDF Code

@inproceedings{bertolissi2025local,
  title = {Local Mixtures of Experts: Essentially Free Test-Time Training via Model Merging},
  author = {Bertolissi, Ryo and Hübotter, Jonas and Hakimi, Ido and Krause, Andreas},
  year = {2025},
  booktitle = {Conference on Language Modeling},
}

ICLR ’25 Best Paper

Efficiently Learning at Test-Time: Active Fine-Tuning of LLMs

Jonas Hübotter, Sascha Bongni, Ido Hakimi , and 1 more author

In International Conference on Learning Representations , 2025

Best Paper Award at NeurIPS Workshop on Fine-Tuning in Modern Machine Learning, 2024.

Bib PDF Code Poster Slides

@inproceedings{hubotter2025efficiently,
  title = {Efficiently Learning at Test-Time: Active Fine-Tuning of LLMs},
  author = {H{\"u}botter, Jonas and Bongni, Sascha and Hakimi, Ido and Krause, Andreas},
  year = {2025},
  booktitle = {International Conference on Learning Representations},
}

NeurIPS ’24 Oral
Transductive Active Learning: Theory and Applications

Jonas Hübotter, Bhavya Sukhija, Lenart Treven , and 2 more authors

In Advances in Neural Information Processing Systems , 2024

Oral Presentation at ICML Workshop on Aligning Reinforcement Learning Experimentalists and Theorists, 2024.

Bib PDF Code Poster Slides
@inproceedings{hubotter2024transductive, title = {Transductive Active Learning: Theory and Applications}, author = {H{\"u}botter, Jonas and Sukhija, Bhavya and Treven, Lenart and As, Yarden and Krause, Andreas}, year = {2024}, booktitle = {Advances in Neural Information Processing Systems}, }

Talks

Towards Solving Hard Problems via Test-Time Training — slides
Invited Talk, 1st Prague Workshop on Neural Networks and Reasoning, Prague, 10 Jul 2025.
Interview on Test-Time Model Merging with Nnamdi Iregbulem (Lightspeed Venture Partners), May 2025.
Efficiently Learning at Test-Time with LLMs via Transductive Active Learning — recording, slides
Invited Talk, Trillion Parameter Consortium (TPC) Seminar Series, 5 Mar 2025.
Efficiently Learning at Test-Time: Active Fine-Tuning of LLMs — recording, slides
Contributed Talk, NeurIPS Workshop on Fine-Tuning in Modern Machine Learning, Vancouver, 14 Dec 2024.
Interview with Machine Learning Street Talk (MLST) podcast, Nov 2024.
Transductive Active Learning for Fine-Tuning Large (Language) Models — slides
Invited Talk, Machine Learning and Modelling Seminar, Czech Academy of Sciences, Prague, 21 Nov 2024.
Efficiently Learning at Test-Time with LLMs — recording, slides
Invited Talk, Zurich AI Meetup, Zurich, 3 Dec 2024.
Invited Talk, Tufa Labs AI Meetup, Zurich, 29 Oct 2024.
Transductive Active Learning with Application to Safe Bayesian Optimization — recording, slides
Contributed Talk, ICML Workshop on Aligning Reinforcement Learning Experimentalists and Theorists, Vienna, 26 Jul 2024.
Active Fine-Tuning of Large Neural Networks — slides
Contributed Talk, Machine Learning Seminar, ETH Zurich, 18 Apr 2024.

Supervision

I have had the privilege of advising several BSc and MSc students during their theses and semester projects. Some of these projects have led to publications.

Matthias Otth: Efficient Fine-Tuning and Test-Time Training of Large Language Models for Reasoning Tasks (with Ido Hakimi, SCALR@COLM '25)
Leander Diaz-Bone: Directed Goal-Conditioned Reinforcement Learning (with Marco Bagatella)
Ryo Bertolissi: Test-Time Model Merging for Mixture of Local Experts (with Ido Hakimi, COLM '25)
Nicolas Menet: Efficiently Estimating Gaussian Probability of Maximality (with Parnian Kassraie, AISTATS '25)
Sascha Bongni: Active Fine-Tuning of Large Language Models (ICLR '25)
Pablo Lahmann: Safe Control as Inference (with Yarden As)
Anh Duc Nguyen: Safe Bayesian Optimization without Regret

You can find a list of potential projects of our research group here. If you want to work with me, please send me an email describing your area of interest. Please also attach your CV and up-to-date transcripts.