22.06.2026

32 papers at ICML 2026 accepted

At this year's ICML conference, 32 papers were accepted from researchers in our Cluster.

The 43rd International Conference on Machine Learning (ICML) is held in Seoul, South Korea from July 6-11, 2026. ICML is the leading conference for Machine Learning. The purpose of the annual meetings is to foster the exchange of research on Machine Learning in its technological, mathematical, and theoretical aspects. The core focus is peer-reviewed novel research which is presented and discussed in the general session, along with invited talks by leaders in their field.

This year our Cluster is represented with 32 papers at ICML.

List of titles of the papers by our members (bold) and their team members (check out the rest of the accepted papers here):

1. Ilze Amanda Auzina, Joschka Strüber, Sergio Hernández-Gutiérrez, Shashwat Goel, Ameya Pandurang Prabhu, Matthias Bethge
   Intrinsic Credit Assignment for Long Horizon Interaction
2. Jana Zeller, Thaddäus Wiedemer, Fanfei Li, Thomas Klein, Prasanna Mayilvahanan, Matthias Bethge, Felix Wichmann, Ryan Cotterell, Wieland Brendel
   MentisOculi: Revealing the Limits of Reasoning with Mental Imagery
3. Ben Rank, Hardik Bhatnagar, Ameya Pandurang Prabhu, Shira Eisenberg, Karina Nguyen, Matthias Bethge, Maksym Andriushchenko
   PostTrainBench: Can LLM Agents Automate LLM Post-Training?
4. Cagatay Yildiz, Nishaanth Kanna, Nitin Sharma, Matthias Bethge, Beyza Ermis
   Investigating Continual Pretraining in Large Language Models: Insights and Implications
5. Shruti Joshi, Aaron Mueller, David Klindt, Wieland Brendel, Dhanya Sridhar, Patrik Reizinger
   Position: Causality is Key for Interpretability Claims to Generalise
6. Jack Brady, Bernhard Schölkopf, Thomas Kipf, Simon Buchholz, Wieland Brendel
   Generation is Required for Data-Efficient Perception
7. Manuel Traub, Martin V Butz
   Looking Locally: Object-Centric Vision Transformers as Foundation Models for Efficient Segmentation
8. Moritz Brösamle, Stephan Eckstein
   The Expressive Power of Low Precision Softmax Transformers with (Summarized) Chain-of-Thought
9. Fabian Morelli, Stephan Eckstein
   Partial Fusion of Neural Networks: Efficient Tradeoffs Between Ensembles and Weight Aggregation
10. Haofei Xu, Rundi Wu, Philipp Henzler, Nikolai Kalischek, Michael Oechsle, Fabian Manhardt, Marc Pollefeys, Andreas Geiger, Federico Tombari, Michael Niemeyer
    PointDiT: Pixel-Space Diffusion for Monocular Geometry Estimation
11. Ricardo Dominguez-Olmedo, Bernhard Schölkopf, Moritz Hardt
    Computational Arbitrage in AI Model Markets
12. Mina Remeli, Moritz Hardt
    Correct looks better: Pairwise comparisons reveal accuracy rankings
13. Yatong Chen, Guanhua Zhang, Moritz Hardt
    Leaderboard Incentives: Model Rankings under Strategic Post-Training
14. Nikhil Chandak, Shashwat Goel, Ameya Pandurang Prabhu, Moritz Hardt, Jonas Geiping
    Curating the Future: A Scalable Recipe for Training Open-Ended Forecasters
15. Tim Weiland, Philipp Hennig
    Scalable Bayesian Inference for Nonlinear Conservation Laws
16. Arik Reuter, Anish Dhir, Cristiana Diaconu, Jake Robertson, Ole Ossen, Frank Hutter, Adrian Weller, Mark van der Wilk, Bernhard Schölkopf
    Use What You Know: Causal Foundation Models with Partial Graphs
17. Herilalaina Rakotoarison, Steven Adriaensen, Tom Viering, Samuel Gabriel Müller, Carl Hvarfner, Frank Hutter, Eytan Bakshy
    $\alpha$-PFN: Fast Entropy Search via In-Context Learning
18. Tom Sühr, Florian Dorner, Olawale Salaudeen, Augustin Kelava, Samira Samadi
    Position: Stop evaluating AI with human tests, develop principled, AI-specific tests instead
19. Han Wang, Weijie Wang, Jiaqi Liu, Hilde Kuehne, Nicu Sebe
    Deep Trajectory Supervision: Deep Supervision Strikes Back
20. Stefan Wahl, Raphaela Schenk, Ali Farnoud, Jakob Macke, Daniel Gedon
    A Probabilistic Framework for LLM-Based Model Discovery
21. Manuel Glöckler, Jose Pedro JP Manzano-Patron, Stamatios Sotiropoulos, Cornelius Schröder, Jakob Macke
    Scalable Simulation-Based Model Inference with Test-Time Complexity Control
22. Mikel Zhobro, Andreas René Geist, Georg Martius
    Learning 3D-Gaussian Simulators from RGB Videos
23. Anselm Paulus, Andreas René Geist, Vit Musil, Sebastian Hoffmann, Georg Martius
    SoftJAX & SoftTorch: Empowering Automatic Differentiation Libraries with Informative Gradients
24. Marco Bagatella, Mert Albaba, Jonas Hübotter, Georg Martius, Andreas Krause
    Test-time Offline Reinforcement Learning on Goal-related Experience
25. Alina Wernick, Kristof Meding
    Position: EU AI Act's Research Exemptions Can Break the Publication Norms of Major AI Conferences
26. Ruta Binkyte, Ivaxi Sheth, Zhijing Jin, Mohammad Havaei, Bernhard Schölkopf, Mario Fritz
    Position: Trustworthy AI Suffers from Invariance Conflicts and Causality is The Solution
27. Sirui Lu, Zhijing Jin, Terry Zhang, Pavel Kos, Juan Cirac, Bernhard Schölkopf
    Position: LLM for Physics Research Requires Domain-Specialized Training and Tooling
28. Samuel Simko, Punya Pandey, Zhijing Jin, Bernhard Schölkopf
    Training with Honeypots: Reshaping How LLMs Fail
29. Sawal Acharya, Terry Zhang, Andrew Kim, Anahita Haghighat, Xianlin Sun, Pepijn Cobben, Rahul Shrestha, Maximilian Mordig, Jacob Emmerson, Furkan Danisman, Yuen Chen, Clijo Jose, Andrei Muresanu, Justin Cui, Jiarui Liu, Yahang Qi, Punya Pandey, Yinya Huang, Bernhard Schölkopf, Zhijing Jin
    CauSciBench: Evaluating LLM Causal Inference for Scientific Research
30. Ziheng Chen, Xiaojun Wu, Bernhard Schölkopf, Nicu Sebe
    Riemannian Networks over Full-Rank Correlation Matrices
31. David Guzman Piedrahita, Dave Banerjee, Changling Li, Terry Zhang, Kevin Blin, Samuel Simko, Punya Pandey, Irene Strauss, Rada Mihalcea, Bernhard Schölkopf, Zhijing Jin
    Position: Safe Models Do Not Guarantee Safe Societies: The Case for Sociopolitical Risk
32. Onno Eberhard, Claire Vernade, Michael Muehlebach
    Commit to the Bit: Reactive Reinforcement Learning Done Right

Share

Back