I am currently a postdoctoral fellow at the Oden Institute, University of Texas at Austin, collaborating with Prof. Umberto Villa on topics related to provable computational methods, interpretable/reliable AI models, and photoacoustic imaging, among others. Prior to Austin, I was a postdoc at the University of Michigan, Ann Arbor, where I collaborated with Prof. Jeffrey A. Fessler and Prof. Luis Hernandez-Garcia on topics related to computational MRI, Arterial Spin Labeling, and inverse problems. I completed my PhD research in the Computer Science Department at the Technion — Israel Institute of Technology, under the supervision of Prof. Irad Yavneh and Prof. Michael Zibulevsky. My graduate research primarily focused on numerical optimization and multigrid computational methods. Thesis link.

My Erdős number is 3.

My research interests include:

• Numerical Optimization & Multigrid Computational Methods
• Interpretable/Reliable AI Models & Scientific Computing & Compressive Sensing
• Photoacoustic Computed Tomography and Computational MR Imaging
• Applications to Machine Learning, Signal Processing, and Computational Imaging

News

2025

Oct 16
Our work Convergent Complex Quasi-Newton Proximal Methods for Gradient-Driven Denoisers in Compressed Sensing MRI Reconstruction was accepted to IEEE Transactions on Computational Imaging! Many thanks to all collaborators. In this paper, we propose a new method with rigorous theoretical guarantees for CS-MRI reconstruction with learned priors. Our theoretical results rely on an assumption that is easily satisfied by practical neural networks. Find our implementation and associated project website at Code and Project Website.

Oct 8
Our work Using Randomized Nyström Preconditioners to Accelerate Variational Image Reconstruction was accepted to IEEE Transactions on Computational Imaging! Many thanks to all collaborators. In this paper, we propose a novel technique for designing an on-the-fly preconditioner using only $\mathrm{A}\mathrm{x}$. Moreover, we demonstrate how this preconditioner can be used to accelerate image reconstruction with wavelet, total variation (TV), and Hessian-Schatten norm regularizers. Find our implementation at Code.

Aug 17
Our latest works A Convergent Generalized Krylov Subspace Method for Compressed Sensing MRI Reconstruction with Gradient-Driven Denoisers (our method can reconstruct a clear image within seconds and comes with a well-established convergence analysis) and A Mini-Batch Quasi-Newton Proximal Method for Constrained Total-Variation Nonlinear Image Reconstruction (we present an approach to estimate approximate Hessian matrices with noisy gradients; experiments on 3D Optical Diffraction Tomography with real data demonstrate the effectiveness of our approach) are now online! Many thanks to all collaborators.

Aug 11 - 15
Attended the IPAM Randomized Numerical Linear Algebra workshop at UCLA. Many thanks to the organizers and our group leaders Xiaoye Sherry Li and Zichao Wendy Di . It was a pleasure to learn and meet so many wonderful colleagues.

Jun 2 - 6
Attended the ICERM Scientific Machine Learning for Gravitational Wave Astronomy workshop at Brown University. Many thanks to the organizers. It was a pleasure to learn more about gravitational waves and meet many new colleagues.

May 08
Our latest work Convergent Complex Quasi-Newton Proximal Methods for Gradient-Driven Denoisers in Compressed Sensing MRI Reconstruction is now online! Many thanks to all collaborators.

2024

Nov 12
Our work Using Randomized Nyström Preconditioners to Accelerate Variational Image Reconstruction is now online! Many thanks to all collaborators.

Oct 01
Our work Provable Preconditioned Plug-and-Play Approach for Compressed Sensing MRI Reconstruction was accepted to IEEE Transactions on Computational Imaging! Many thanks to all collaborators. (Poster) (Slides) (PDF)

Jan 26
Our work A Complex Quasi-Newton Proximal Method for Image Reconstruction in Compressed Sensing MRI was accepted to IEEE Transactions on Computational Imaging! Many thanks to all collaborators. (Slides) (Code) (PDF)

Contact Information:

Oden Institute
University of Texas at Austin
TX, 78712-1229, USA
Email: tao + . + lastname@austin.utexas.edu