Jonas Adler is a Senior Research Scientist at Google DeepMind, where he works on scientific machine learning, focusing on inverse problems and the intersection between model-driven and data-driven methods.

Adler pursued a PhD in Applied Mathematics under the supervision of Ozan Öktem before working as a Research Scientist at Elekta. He joined the Science team at DeepMind in June 2019.

Adler has authored and co-authored numerous papers, including:

- "Learning to solve inverse problems using Wasserstein loss" (2017)
- "Model based learning for accelerated, limited-view 3D photoacoustic tomography" (2017)
- "Learned Primal-Dual Reconstruction" (2017)
- "GPUMCI: a flexible platform for x-ray imaging on the GPU" (2017)
- "Solving ill-posed inverse problems using iterative deep neural networks" (2017)
- "Spectral CT reconstruction with anti-correlated noise model and joint prior" (2017)
- "A modified fuzzy C means algorithm for shading correction in craniofacial CBCT images" (2017)
- "Task adapted reconstruction for inverse problems" (2018)
- "Deep Bayesian Inversion" (2018)
- "Deep Learning Framework for Digital Breast Tomosynthesis Reconstruction" (2018)
- "Data-driven Nonsmooth Optimization" (2018)
- "EDS tomographic reconstruction regularized by total nuclear variation joined with HAADF-STEM tomography" (2018)
- "Banach Wasserstein GAN" (2018)
- "On the unreasonable effectiveness of CNNs" (2020)
- "A unified representation network for segmentation with missing modalities" (2019)
- "Multi-Scale Learned Iterative Reconstruction" (2019)
- "Inferring a Continuous Distribution of Atom Coordinates from Cryo-EM Images using VAEs" (co-authored, 2021)
- "Continuous diffusion for categorical data" (co-authored, 2022)

Jonas Adler is a Senior Research Scientist at Google DeepMind, where he works on scientific machine learning, focusing on inverse problems and the intersection between model-driven and data-driven methods.

Adler pursued a PhD in Applied Mathematics under the supervision of Ozan Öktem before becoming a Research Scientist at Elekta. He joined the Science team at DeepMind in June 2019.

Adler's research interests lie at the intersection of machine learning and the natural sciences. He has worked on AlphaFold, recognised as a solution to the protein folding problem at CASP14. He has also published extensively on image reconstruction tasks, including the following:

- Learned Primal-Dual Reconstruction
- Banach Wasserstein GAN
- Deep Bayesian Inversion
- Task adapted reconstruction for inverse problems
- Deep Learning Framework for Digital Breast Tomosynthesis Reconstruction
- Data-driven Nonsmooth Optimisation
- EDS tomographic reconstruction regularised by total nuclear variation joined with HAADF-STEM tomography
- Learning to solve inverse problems using Wasserstein loss
- Model-based learning for accelerated, limited-view 3D photoacoustic tomography

- Andreas Hauptmann, Jonas Adler. On the unreasonable effectiveness of CNNs. 2020
- Nikita Moriakov, Jonas Adler, Jonas Teuwen. Kernel of CycleGAN as a Principle homogeneous space. 2020
- Kenneth Lau, Jonas Adler, Jens Sjölund. A unified representation network for segmentation with missing modalities. 2019
- Andreas Hauptmann, Jonas Adler, et al. Multi-Scale Learned Iterative Reconstruction. 2019
- Jonas Adler, Ozan Öktem. Deep Bayesian Inversion. 2018
- Jonas Adler, et al. Task adapted reconstruction for inverse problems. 2018
- Nikita Moriakov, et al. Deep Learning Framework for Digital Breast Tomosynthesis Reconstruction. 2018
- Sebastian Banert, et al. Data-driven Nonsmooth Optimization. 2018
- Zhichao Zhong, et al. EDS tomographic reconstruction regularized by total nuclear variation joined with HAADF-STEM tomography. 2018
- Jonas Adler, Sebastian Lunz. Banach Wasserstein GAN. 2018
- Jonas Adler, et al. Learning to solve inverse problems using Wasserstein loss. 2017
- Andreas Hauptmann, et al. Model based learning for accelerated, limited-view 3D photoacoustic tomography. 2017
- Jonas Adler, Ozan Öktem. Learned Primal-Dual Reconstruction. 2017
- Jonas Adler, Gregory J. Bootsma, et al. GPUMCI: a flexible platform for x-ray imaging on the GPU. 2017
- Jonas Adler, Ozan Öktem. Solving ill-posed inverse problems using iterative deep neural networks. 2017
- Mats Persson, Jonas Adler. Spectral CT reconstruction with anti-correlated noise model and joint prior. 2017

Jonas Adler is a Senior Research Scientist at Google DeepMind, where he works on scientific machine learning, focusing on inverse problems and the intersection between model-driven and data-driven methods.

Adler completed his PhD in Applied Mathematics under the supervision of Ozan Öktem before becoming a Research Scientist at Elekta. He then moved to DeepMind in 2019, joining the Science team.

Adler's research sits at the intersection of machine learning and the natural sciences, with a particular focus on image reconstruction and inverse problems. He has worked on AlphaFold, recognised as a solution to the protein folding problem.

Adler has published extensively, including the following:

- "Accelerated Forward-Backward Optimisation using Deep Learning" (arXiv, 2021)
- "Deep Bayesian Inversion" (arXiv, 2018)
- "Task adapted reconstruction for inverse problems" (arXiv, 2018)
- "Deep Learning Framework for Digital Breast Tomosynthesis Reconstruction" (arXiv, 2018)
- "Data-driven Nonsmooth Optimisation" (arXiv, 2018)
- "EDS tomographic reconstruction regularized by total nuclear variation joined with HAADF-STEM tomography" (Ultramicroscopy, 2018)
- "Banach Wasserstein GAN" (NIPS, 2018)
- "Learning to solve inverse problems using Wasserstein loss" (NIPS workshop in optimal transport, 2017)
- "Model based learning for accelerated, limited-view 3D photoacoustic tomography" (IEEE - Transactions on Medical Imaging, 2017)
- "Learned Primal-Dual Reconstruction" (IEEE - Transactions on Medical Imaging, 2017)
- "GPUMCI: a flexible platform for x-ray imaging on the GPU" (Fully3D, 2017)
- "Solving ill-posed inverse problems using iterative deep neural networks" (Inverse Problems, 2017)

sness@sness.net