Thomas Möllenhoff

Research Associate
RIKEN Center for Advanced Intelligence Project
Approximate Bayesian Inference Team
Tokyo, Japan

thomas.moellenhoff [at] riken [.] jp
CVScholargithubTwitter



Since August 2020, I'm a researcher at the RIKEN Center for Advanced Intelligence Project and a member of the Approximate Bayesian Inference Team, where I work with Emtiyaz Khan. Before that, I did my PhD in the Computer Vision Group at the Technical University of Munich under the guidance of Daniel Cremers.

My current research focuses on the design and analysis of data-driven algorithms using ideas from mathematical optimization. My long-term goal is to discover fundamental principles behind robust and efficient machine learning systems, enabling them to continually learn and adapt in an uncertain and dynamic environment.

I have been working on problems connected to the following research areas:

Publications

2020

Z. Ye, T. Möllenhoff, T. Wu, D. Cremers. Optimization of Graph Total Variation via Active-Set-based Combinatorial Reconditioning. In Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS), 2020. [pdf] [code]

2019

P. Bréchet, T. Wu, T. Möllenhoff, D. Cremers. Informative GANs via structured regularization of optimal transport. Optimal Transport and Machine Learning (NeurIPS Workshop), 2019. [pdf]

T. Möllenhoff, D. Cremers. Lifting vectorial variational problems: A natural formulation based on geometric measure theory and discrete exterior calculus. In Proceedings of the Conference on Computer Vision and Pattern Recognition (CVPR), 2019. [pdf] [talk]

T. Möllenhoff, D. Cremers. Flat metric minimization with applications in generative modeling. In Proceedings of the International Conference on Machine Learning (ICML), 2019. [pdf] [code] [talk] [poster]

M. Moeller, T. Möllenhoff, D. Cremers. Controlling neural networks via energy dissipation. In Proceedings of the International Conference on Computer Vision (ICCV), 2019. [pdf]

2018

B. Haefner, T. Möllenhoff, Y. Queau, D. Cremers. Fight ill-posedness with ill-posedness: Single-shot variational depth super-resolution from shading. In Proceedings of the Conference on Computer Vision and Pattern Recognition (CVPR), 2018. [pdf] [code]

T. Frerix*, T. Möllenhoff*, M. Moeller*, D. Cremers. Proximal backpropagation. In Proceedings of the International Conference on Learning Representations (ICLR), 2018. [pdf] [code]

T. Möllenhoff, Z. Ye, T. Wu, D. Cremers. Combinatorial preconditioners for proximal algorithms on graphs. In Proceedings of the International Conference on Artificial Intelligence and Statistics (AISTATS), 2018. [pdf]

2017

T. Möllenhoff, D. Cremers. Sublabel-accurate discretization of nonconvex free-discontinuity problems. In Proceedings of the International Conference on Computer Vision (ICCV), 2017. [pdf]

2016

E. Laude*, T. Möllenhoff*, M. Moeller, J. Lellmann, D. Cremers. Sublabel-accurate convex relaxation of vectorial multilabel energies. In Proceedings of the European Conference on Computer Vision (ECCV), 2016. [pdf] [code]

T. Möllenhoff*, E. Laude*, M. Moeller, J. Lellmann, D. Cremers. Sublabel-accurate relaxation of nonconvex energies. In Proceedings of the Conference on Computer Vision and Pattern Recognition (CVPR), 2016. [pdf] [code] [talk]

2015

T. Möllenhoff, E. Strekalovskiy, M. Moeller, D. Cremers. The primal-dual hybrid gradient method for semiconvex splittings. SIAM Journal on Imaging Sciences, 2015. [pdf] [talk] [slides]

T. Möllenhoff, E. Strekalovskiy, M. Moeller, D. Cremers. Low rank priors for color image regularization. International Workshop on Energy Minimization Methods in Computer Vision and Pattern Recognition (EMMCVPR), 2015. [pdf]

2013

T. Möllenhoff, E. Toeppe, C. Nieuwenhuis, D. Cremers. Efficient convex optimization for minimal partition problems with volume constraints. International Workshop on Energy Minimization Methods in Computer Vision and Pattern Recognition (EMMCVPR), 2013. [pdf]


* contribution.

Code