publications

2025

1. GradPCA: Leveraging NTK Alignment for Reliable Out-of-Distribution Detection

   Mariia Seleznova, Hung-Hsu Chou, Claudio Mayrink Verdun, and Gitta Kutyniok

   arXiv preprint arXiv:2505.16017, 2025
2. Multi-Group Proportional Representations for Text-to-Image Models

   Sangwon Jung, Alex Oesterling, Claudio Mayrink Verdun, Sajani Vithana, and 2 more authors

   In CVPR, 2025
3. Get rid of your constraints and reparametrize: A study in NNLS and implicit bias

   Hung-Hsu Chou, Johannes Maly, Claudio Mayrink Verdun, Bernardo Freitas Paulo Costa, and 1 more author

   In AISTATS, 2025

4. 

   Inference-Time Reward Hacking in Large Language Models

   Hadi Khalaf, Claudio Mayrink Verdun, Alex Oesterling, Himabindu Lakkaraju, and 1 more author

   NeurIPS 2025 (Spotlight, top 3%), 2025
5. ProofCompass: Enhancing Specialized Provers with LLM Guidance

   Nicolas Wischermann, Claudio Mayrink Verdun, Gabriel Poesia, and Francesco Noseda

   In 2nd AI for Math Workshop@ ICML 2025, 2025
6. HeavyWater and SimplexWater: Watermarking Low-Entropy Text Distributions

   Dor Tsur, Carol Xuan Long, Claudio Mayrink Verdun, Hsiang Hsu, and 4 more authors

   NeurIPS, 2025

7. 

   AI Alignment at Your Discretion

   Maarten Buyl, Hadi Khalaf, Claudio Mayrink Verdun, Lucas Monteiro Paes, and 2 more authors

   In ACM FAccT (and Best Paper Award at New England NLP Symposium), 2025

8. 

   Optimized Couplings for Watermarking Large Language Models

   Carol Xuan Long, Dor Tsur, Claudio Mayrink Verdun, Hsiang Hsu, and 2 more authors

   In IEEE International Symposium on Information Theory (ISIT), 2025
9. Leveraging the Sequential Nature of Language for Interpretability

   Usha Bhalla, Alex Oesterling, Claudio Mayrink Verdun, Flavio P. Calmon, and 1 more author

   In Workshop on Assessing World Models, ICML (Spotlight), 2025

2024

1. Non-Asymptotic Uncertainty Quantification in High-Dimensional Learning

   Frederik Hoppe, Claudio Mayrink Verdun, Hannah Laus, Felix Krahmer, and 1 more author

   NeurIPS (Spotlight, top 2%), 2024

   Bib HTML

   @article{hoppe2024non,
     title = {Non-Asymptotic Uncertainty Quantification in High-Dimensional Learning},
     author = {Hoppe, Frederik and Verdun, Claudio Mayrink and Laus, Hannah and Krahmer, Felix and Rauhut, Holger},
     journal = {NeurIPS (Spotlight, top 2%)},
     volume = {34},
     pages = {2873--2886},
     year = {2024},
   }

2. 

   Measuring progress in dictionary learning for language model interpretability with board game models

   Adam Karvonen, Benjamin Wright, Can Rager, Rico Angell, and 5 more authors

   NeurIPS, 2024

   Bib HTML

   @article{karvonen2024measuring,
     title = {Measuring progress in dictionary learning for language model interpretability with board game models},
     author = {Karvonen, Adam and Wright, Benjamin and Rager, Can and Angell, Rico and Brinkmann, Jannik and Smith, Logan and Verdun, Claudio Mayrink and Bau, David and Marks, Samuel},
     journal = {NeurIPS},
     volume = {34},
     pages = {2873--2886},
     year = {2024},
   }

3. 

   Multi-Group Proportional Representation

   Alex Oesterling, Claudio Mayrink Verdun, Carol Xuan Long, Alex Glynn, and 4 more authors

   NeurIPS, 2024

   Bib HTML

   @article{oesterling2024multi,
     title = {Multi-Group Proportional Representation},
     author = {Oesterling, Alex and Mayrink Verdun, Claudio and Long, Carol Xuan and Glynn, Alex and Paes, Lucas Monteiro and Vithana, Sajani and Cardone, Martina and Calmon, Flavio P},
     journal = {NeurIPS},
     volume = {34},
     pages = {2873--2886},
     year = {2024},
   }

4. Imaging with Confidence: Uncertainty Quantification for High-dimensional Undersampled MR Images

   Frederik Hoppe, Claudio Mayrink Verdun, Hannah Laus, Sebastian Endt, and 3 more authors

   In ECCV, 2024

   Bib HTML

   @inproceedings{hoppe2024confidence,
     title = {Imaging with Confidence: Uncertainty Quantification for High-dimensional Undersampled MR Images},
     author = {Hoppe, Frederik and Mayrink Verdun, Claudio and Laus, Hannah and Endt, Sebastian and Menzel, Marion and Krahmer, Felix and Rauhut, Holger},
     booktitle = {ECCV},
     pages = {1--19},
     year = {2024},
     organization = {Springer International Publishing}
   }

5. Fast, blind, and accurate: Tunning-free sparse regression with global linear convergence

   Claudio Mayrink Verdun, Oleh Melnyk, Felix Krahmer, and Peter Jung

   In COLT, 2024

   Bib HTML

   @inproceedings{jung2024regression,
     title = {Fast, blind, and accurate: Tunning-free sparse regression with global linear convergence},
     author = {Mayrink Verdun, Claudio and Melnyk, Oleh and Krahmer, Felix and Jung, Peter},
     booktitle = {COLT},
     pages = {1--45},
     year = {2024},
     organization = {IEEE}
   }

6. With or Without Replacement? Improving Confidence in Fourier Imaging

   Frederik Hoppe, Claudio Mayrink Verdun, Felix Krahmer, Marion Menzel, and 1 more author

   In COSERA, 2024

   Bib HTML

   @inproceedings{hoppe2024replacement,
     title = {With or Without Replacement? Improving Confidence in Fourier Imaging},
     author = {Hoppe, Frederik and Mayrink Verdun, Claudio and Krahmer, Felix and Menzel, Marion and Rauhut, Holger},
     booktitle = {COSERA},
     pages = {1--45},
     year = {2024},
     organization = {IEEE}
   }

7. AI Technologies: Algorithmic Monocultures, Arbitrariness, and Global Divides

   Claudio Mayrink Verdun, Camila A Tsuzuki, Caio V Machado, Juan Felipe Gomez, and 2 more authors

   In T20 conference, 2024

   Bib HTML

   @inproceedings{mayrinkverdun2024monocultures,
     title = {AI Technologies: Algorithmic Monocultures, Arbitrariness, and Global Divides},
     author = {Mayrink Verdun, Claudio and Tsuzuki, Camila A and Machado, Caio V and Gomez, Juan Felipe and Monteiro Paes, Lucas and Calmon, Flavio P},
     booktitle = {T20 conference},
     pages = {1--45},
     year = {2024},
     organization = {G20 summit}
   }

2023

1. Uncertainty quantification for learned ISTA

   Frederik Hoppe, Claudio Mayrink Verdun, Hannah Laus, Felix Krahmer, and 1 more author

   In MLSP, 2023

   Bib HTML

   @inproceedings{hoppe2023uncertainty,
     title = {Uncertainty quantification for learned ISTA},
     author = {Hoppe, Frederik and Mayrink Verdun, Claudio and Laus, Hannah and Krahmer, Felix and Rauhut, Holger},
     booktitle = {MLSP},
     pages = {1--6},
     year = {2023},
     organization = {IEEE}
   }

2. 

   In-vivo myelin water quantification using diffusion-relaxation correlation MRI: a comparison of 1D and 2D methods

   Sebastian Endt, Maria Engel, Emanuele Naldi, Rodolfo Assereto, and 7 more authors

   Applied Magnetic Resonance, 2023

   Bib HTML

   @article{endt2023vivo,
     title = {In-vivo myelin water quantification using diffusion-relaxation correlation MRI: a comparison of 1D and 2D methods},
     author = {Endt, Sebastian and Engel, Maria and Naldi, Emanuele and Assereto, Rodolfo and Molendowska, Malwina and M{\"u}ller, Lars and Mayrink Verdun, Claudio and Pirkl, Carolin M. and Palombo, Marco and Jones, DereK K. and Menzel, Marion I.},
     journal = {Applied Magnetic Resonance},
     volume = {54},
     pages = {1571–1588},
     year = {2023}
   }

3. Sampling Strategies for Compressive Imaging Under Statistical Noise

   Frederik Hoppe, Felix Krahmer, Claudio Mayrink Verdun, Marion I Menzel, and 1 more author

   In SampTA 2023, 2023

   HTML

4. 

   High-Dimensional Confidence Regions in Sparse MRI

   Frederik Hoppe, Felix Krahmer, Claudio Mayrink Verdun, Marion I. Menzel, and 1 more author

   In ICASSP (Best Student Paper Award), 2023

   Bib HTML

   @inproceedings{hoppe2023high,
     title = {High-Dimensional Confidence Regions in Sparse MRI},
     author = {Hoppe, Frederik and Krahmer, Felix and Mayrink Verdun, Claudio and Menzel, Marion I. and Rauhut, Holger},
     booktitle = {ICASSP (Best Student Paper Award)},
     pages = {1--5},
     year = {2023},
     organization = {IEEE},
   }

2022

1. Uncertainty quantification for sparse Fourier recovery

   Frederik Hoppe, Felix Krahmer, Claudio Mayrink Verdun, Marion I Menzel, and 1 more author

   arXiv preprint arXiv:2212.14864, 2022

   Bib HTML

   @article{hoppe2022uncertainty,
     title = {Uncertainty quantification for sparse Fourier recovery},
     author = {Hoppe, Frederik and Krahmer, Felix and Verdun, Claudio Mayrink and Menzel, Marion I and Rauhut, Holger},
     journal = {arXiv preprint arXiv:2212.14864},
     year = {2022}
   }

2. Non-negative least squares via overparametrization

   Hung-Hsu Chou, Johannes Maly, and Claudio Mayrink Verdun

   arXiv preprint arXiv:2207.08437, 2022

   Bib HTML

   @article{chou2022non,
     title = {Non-negative least squares via overparametrization},
     author = {Chou, Hung-Hsu and Maly, Johannes and Mayrink Verdun, Claudio},
     journal = {arXiv preprint arXiv:2207.08437},
     year = {2022}
   }

2021

1. Unmixing tissue compartments via deep learning T1-T2-relaxation correlation imaging

   Sebastian Endt, Carolin M Pirkl, Claudio M Verdun, Bjoern H Menze, and 1 more author

   In 17th International Symposium on Medical Information Processing and Analysis, 2021

2. 

   A scalable second order method for ill-conditioned matrix completion from few samples

   Christian Kümmerle, and Claudio Mayrink Verdun

   In ICML (Spotlight), 2021

   Bib HTML

   @inproceedings{kummerle2021scalable,
     title = {A scalable second order method for ill-conditioned matrix completion from few samples},
     author = {K{\"u}mmerle, Christian and Mayrink Verdun, Claudio},
     booktitle = {ICML (Spotlight)},
     pages = {5872--5883},
     year = {2021},
     organization = {PMLR},
   }

3. 

   Group testing for SARS-CoV-2 allows for up to 10-fold efficiency increase across realistic scenarios and testing strategies

   Claudio Mayrink Verdun, Tim Fuchs, Pavol Harar, Dennis Elbrächter, and 5 more authors

   Frontiers in Public Health (highlighted by David Donoho at https://www.youtube.com/watch?v=VOzl-RC4IIs, 2021

   Bib HTML

   @article{verdun2021group,
     title = {Group testing for SARS-CoV-2 allows for up to 10-fold efficiency increase across realistic scenarios and testing strategies},
     author = {Mayrink Verdun, Claudio and Fuchs, Tim and Harar, Pavol and Elbr{\"a}chter, Dennis and Fischer, David S and Berner, Julius and Grohs, Philipp and Theis, Fabian J and Krahmer, Felix},
     journal = {Frontiers in Public Health (highlighted by David Donoho at https://www.youtube.com/watch?v=VOzl-RC4IIs},
     volume = {9},
     pages = {583377},
     year = {2021},
     publisher = {Frontiers Media SA},
   }

4. 

   Iteratively reweighted least squares for basis pursuit with global linear convergence rate

   Christian Kümmerle, Claudio Mayrink Verdun, and Dominik Stöger

   NeurIPS (Spotlight, top 3%), 2021

   Bib HTML

   @article{kummerle2021iteratively,
     title = {Iteratively reweighted least squares for basis pursuit with global linear convergence rate},
     author = {K{\"u}mmerle, Christian and Mayrink Verdun, Claudio and St{\"o}ger, Dominik},
     journal = {NeurIPS (Spotlight, top 3%)},
     volume = {34},
     pages = {2873--2886},
     year = {2021},
   }

2020

1. Escaping Saddle Points in Ill-Conditioned Matrix Completion with a Scalable Second Order Method

   Christian Kümmerle, and Claudio Mayrink Verdun

   In ICML Workshop - “Beyond first-order methods in ML systems”, 2020

   Bib

   @inproceedings{kuemmerle_verdun,
     author = {K\"ummerle, Christian and Mayrink Verdun, Claudio},
     booktitle = {ICML Workshop - ``Beyond first-order methods in ML systems''},
     pages = {},
     title = {Escaping Saddle Points in Ill-Conditioned Matrix Completion with a Scalable Second Order Method},
     year = {2020}
   }

2019

1. Completion of structured low-rank matrices via iteratively reweighted least squares

   Christian Kümmerle, and Claudio M Verdun

   In SampTA, 2019

   HTML