Jonas Hübotter

• Ph.D. Student
• jonas.huebotter@inf.ethz.ch
• OAT Y 17
• Linkedln
• External Website
• My current 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.

---

Publications

2026

Reinforcement Learning via Self-Distillation J. Hübotter, F. Lübeck*, L. Behric*, A. Baumann*, M. Bagatella, D. Marta, I. Hakimi, I. Shenfeld, T. Kleine Buening, C. Guestrin, A. Krause In International Conference on Machine Learning (ICML), 2026 Best paper award at Test-Time Updates Workshop and oral presentation at Scaling Post-Training for LLMs Workshop at ICLR [bibtex] [abstract] [pdf]

Test-time Offline Reinforcement Learning on Goal-related Experience M. Bagatella*, M. Albaba*, J. Hübotter, G. Martius, A. Krause In International Conference on Machine Learning (ICML), 2026 [bibtex] [abstract] [pdf]

Specialization after Generalization: Towards Understanding Test-Time Training in Foundation Models J. Hübotter*, P. Wolf*, A. Shevchenko*, D. Jüni, A. Krause, G. Kur In International Conference on Learning Representations (ICLR), 2026 Oral presentation at Continual and Compatible Foundation Model Updates Workshop at NeurIPS [bibtex] [abstract] [pdf]

2025

DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning L. Diaz-Bone*, M. Bagatella*, J. Hübotter*, A. Krause In Advances in Neural Information Processing Systems (NeurIPS), 2025 [bibtex] [abstract] [pdf]

Local Mixtures of Experts: Essentially Free Test-Time Training via Model Merging R. Bertolissi*, J. Hübotter*, I. Hakimi, A. Krause In Proc. Conference on Language Modeling (COLM), 2025 [bibtex] [abstract] [pdf]

Active Fine-Tuning of Multi-Task Policies M. Bagatella, J. Hübotter, G. Martius, A. Krause In Proc. International Conference on Machine Learning (ICML), 2025 [bibtex] [abstract] [pdf]

LITE: Efficiently Estimating Gaussian Probability of Maximality N. Menet, J. Hübotter, P. Kassraie, A. Krause In Proc. International Conference on Artificial Intelligence and Statistics (AISTATS), 2025 [bibtex] [abstract] [pdf]

Efficiently Learning at Test-Time: Active Fine-Tuning of LLMs J. Hübotter, S. Bongni, I. Hakimi, A. Krause In Proc. International Conference on Learning Representations (ICLR), 2025 Best paper award at Fine-Tuning in Modern Machine Learning Workshop at NeurIPS [bibtex] [abstract] [pdf]

2024

Transductive Active Learning: Theory and Applications J. Hübotter, B. Sukhija, L. Treven, Y. As, A. Krause In Proc. Neural Information Processing Systems (NeurIPS), 2024 Oral presentation at Aligning Reinforcement Learning Experimentalists and Theorists Workshop at ICML [bibtex] [abstract] [pdf]

2023

Efficient Exploration in Continuous-time Model-based Reinforcement Learning L. Treven, J. Hübotter, B. Sukhija, F. Dörfler, A. Krause In Proc. Neural Information Processing Systems (NeurIPS), 2023 [bibtex] [abstract] [pdf]

Tuning Legged Locomotion Controllers via Safe Bayesian Optimization D. Widmer, D. Kang, B. Sukhija, J. Hübotter, A. Krause, S. Coros In Proc. Conference on Robot Learning (CoRL), 2023 [bibtex] [abstract] [pdf]