Ali Siahkoohi

Ali Siahkoohi

Simons Postdoctoral Fellow
Department of Computational and Applied Mathematics
Rice University

alisk@rice.edu
CV
GitHub
Google Scholar

Research interests

I conduct multidisciplinary research under the guidance of Maarten de Hoop and Richard Baraniuk that focuses on developing deep learning methods to reliably solve scientific computing problems pertaining to inverse problems, uncertainty quantification, and signal processing.

Keywords: Deep Learning, Generative Models, Variational Inference, Inverse Problems, Uncertainty Quantification, Signal Processing

Education

PhD, Georgia Institute of Technology
Computational Science and Engineering

MSc, University of Tehran
Geophysics

BSc, Sharif University of Technology
Electrical Engineering

Publications and presentations

Ali Siahkoohi, Michael Chinen, Tom Denton, W. Bastiaan Kleijn, and Jan Skoglun. Ultra-Low-Bitrate Speech Coding with Pretrained Transformers. In: Proceedings of Interspeech 2022. Sept. 2022, pp. 4421–4425. doi: 10.21437/Interspeech.2022-10988.
PDF Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, and Felix J. Herrmann. Deep Bayesian inference for seismic imaging with tasks. In: Geophysics 87.5. Sept. 2022, pp. S281–S302. doi: 10.1190/geo2021-0666.1.
PDF Code Link BibTeX

Rafael Orozco, Ali Siahkoohi, Gabrio Rizzuti, Tristan van Leeuwen, and Felix J. Herrmann. Adjoint operators enable fast and amortized machine learning based Bayesian uncertainty quantification. Aug. 2022.
PDF Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, Rafael Orozco, and Felix J. Herrmann. Reliable amortized variational inference with conditional normalizing flows via physics-based latent distribution correction. Presented at IMAGE Workshop on Subsurface Uncertainty Description and Estimation - Moving Away from Single Prediction with Distribution Learning. Aug. 2022.
Slides Link BibTeX

Ali Siahkoohi, Mathias Louboutin, and Felix J. Herrmann. Velocity continuation with Fourier neural operators for accelerated uncertainty quantification. In: SEG Technical Program Expanded Abstracts. Aug. 2022, pp. 1765–1769. doi: 10.1190/image2022-3750475.1.
PDF Slides Code Link BibTeX

Mathias Louboutin, Philipp Witte, Ali Siahkoohi, Gabrio Rizzuti, Ziyi Yin, Rafael Orozco, and Felix J. Herrmann. Accelerating innovation with software abstractions for scalable computational geophysics. In: SEG Technical Program Expanded Abstracts. Aug. 2022, pp. 1482–1486. doi: 10.1190/image2022-3750561.1.
PDF Slides Link BibTeX

Ziyi Yin, Ali Siahkoohi, Mathias Louboutin, and Felix J. Herrmann. Learned coupled inversion for carbon sequestration monitoring and forecasting with Fourier neural operators. In: SEG Technical Program Expanded Abstracts. Aug. 2022, pp. 467–472. doi: 10.1190/image2022-3722848.1.
PDF Slides Code Link BibTeX

Yijun Zhang, Mathias Louboutin , Ali Siahkoohi, Ziyi Yin, Rajiv Kumar, and Felix J. Herrmann. A simulation-free seismic survey design by maximizing the spectral gap. In: SEG Technical Program Expanded Abstracts. Aug. 2022, pp. 15–20. doi: 10.1190/image2022-3751690.1.
PDF Slides Code Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, Rafael Orozco, and Felix J. Herrmann. Reliable amortized variational inference with physics-based latent distribution correction. July 2022.
PDF Slides Code Link BibTeX

Ali Siahkoohi. Deep generative models for solving geophysical inverse problems. PhD Thesis, Georgia Institute of Technology.
PDF Slides Link BibTeX

Ali Siahkoohi, Thomas J. Grady II, Abhinav P. Gahlot, Huseyin Tuna Erdinc, and Felix J. Herrmann. Capturing velocity-model uncertainty and two-phase flow with Fourier Neural Operators. Presented at EAGE AI in Geoscience and Geophysics: Current Trends and Future Prospects Dedicated Session. June 2022.
Slides Link BibTeX

Ali Siahkoohi, Rafael Orozco, Gabrio Rizzuti, and Felix J. Herrmann. Wave-equation based inversion with amortized variational Bayesian inference. Presented at EAGE Deep learning for seismic processing: Investigating the foundations workshop. June 2022.
PDF Slides Code Link BibTeX

Ziyi Yin, Mathias Louboutin, Philipp A. Witte, Ali Siahkoohi, Gabrio Rizzuti, Rafael Orozco, Henryk Modzelewski, and Felix J. Herrmann. Julia for Geoscience. Presented at Transform. Mar. 2022.
Video Code Link BibTeX

Mathias Louboutin, Ali Siahkoohi, Ziyi Yin, Rafael Orozco, Thomas J. Grady II, Yijun Zhang, Philipp A. Witte, Gabrio Rizzuti, and Felix J. Herrmann Abstractions for at-scale seismic inversion. Presented at Rice Oil and Gas High Performance Computing Conference. Mar. 2022.
Slides Video Link BibTeX

Yuxiao Ren, Philipp A. Witte, Ali Siahkoohi, Mathias Louboutin, Ziyi Yin, and Felix J. Herrmann. Seismic Velocity Inversion and Uncertainty Quantification Using Conditional Normalizing Flows. Presented at American Geophysical Union (AGU) Fall Meeting. Dec. 2021.
Link BibTeX

Rafael Orozco, Ali Siahkoohi, Gabrio Rizzuti, and Felix J. Herrmann. Variational inference for artifact removal of adjoint solutions in photoacoustic problems. Presented at ML4Seismic Partners Meeting. Nov. 2021.
Slides Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, and Felix J. Herrmann. Uncertainty quantification in imaging and automatic horizon tracking—a Bayesian deep-prior based approach. Presented at ML4Seismic Partners Meeting. Nov. 2021.
Slides Link BibTeX

Ali Siahkoohi, Rafael Orozco, Gabrio Rizzuti, Philipp A. Witte, Mathias Louboutin, and Felix J. Herrmann. Multifidelity conditional normalizing flows for physics-guided Bayesian inference. Presented at ML4Seismic Partners Meeting. Nov. 2021.
Slides Link BibTeX

Rafael Orozco, Ali Siahkoohi, Gabrio Rizzuti, Tristan van Leeuwen, and Felix J. Herrmann. Photoacoustic imaging with conditional priors from normalizing flows. In: Presented at NeurIPS 2021 Workshop on Deep Learning and Inverse Problems. Dec. 2021.
PDF Poster Link

Felix J. Herrmann, Ali Siahkoohi, Rafael Orozco, Gabrio Rizzuti, Philipp A. Witte, and Mathias Louboutin. Learned wave-based imaging—variational inference at scale. Presented at Delft. June 2021.
Slides Link BibTeX

Ali Siahkoohi, Rafael Orozco, Gabrio Rizzuti, Philipp A. Witte, Mathias Louboutin, and Felix J. Herrmann. Fast and reliability-aware seismic imaging with conditional normalizing flows. Presented at Intelligent illumination of the Earth Workshop. June 2021.
Slides Link BibTeX

Mathias Louboutin, Ali Siahkoohi, Rongrong Wang, and Felix J. Herrmann. Low-memory stochastic backpropagation with multi-channel randomized trace estimation. June 2021.
PDF Code Link BibTeX

Ali Siahkoohi and Felix J. Herrmann. Learning by example: fast reliability-aware seismic imaging with normalizing flows. In: SEG Technical Program Expanded Abstracts. Sept. 2021, pp. 1580–1585. doi: 10.1190/segam2021-3581836.1.
PDF Slides Code Link BibTeX

Rajiv Kumar, Maria Kotsi, Ali Siahkoohi, and Alison Malcolm. Enabling uncertainty quantification for seismic data pre-processing using normalizing flows (NF)—an interpolation example. In: SEG Technical Program Expanded Abstracts. Sept. 2021, pp. 1515–1519. doi: 10.1190/segam2021-3583705.1.
PDF Code Link BibTeX

Felix J. Herrmann, Mathias Louboutin, and Ali Siahkoohi. ML@scale using randomized linear algebra. Presented at Microsoft. Mar. 2021.
Slides Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, Mathias Louboutin, Philipp A. Witte, and Felix J. Herrmann. Deep Bayesian Inference for Task-based Seismic Imaging. Presented at KAUST. Mar. 2021.
Slides Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, Mathias Louboutin, Philipp A. Witte, and Felix J. Herrmann. Preconditioned training of normalizing flows for variational inference in inverse problems. In: 3rd Symposium on Advances in Approximate Bayesian Inference. Jan. 2021.
PDF Slides Code Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, Mathias Louboutin, and Felix J. Herrmann. Unsupervised data-guided uncertainty analysis in imaging and horizon tracking. In: 3rd Annual Meeting of the SIAM Texas-Louisiana Section. Oct. 2020.
Slides Link BibTeX

Ali Siahkoohi, Philipp A. Witte, Mathias Louboutin, Felix J. Herrmann, and Gabrio Rizzuti. Seismic Imaging with Uncertainty Quantification: Sampling from the Posterior with Generative Networks. In: SIAM Conference on Imaging Science. IS20. July 2020.
Slides Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, Philipp A. Witte, and Felix J. Herrmann Faster Uncertainty Quantification for Inverse Problems with Conditional Normalizing Flows. In: Tech. rep. TR-CSE-2020-2, Georgia Institute of Technology. July 2020.
PDF Link BibTeX

Gabrio Rizzuti, Ali Siahkoohi, Philipp A. Witte, and Felix J. Herrmann. Parameterizing uncertainty by deep invertible networks, an application to reservoir characterization. In: SEG Technical Program Expanded Abstracts. Sept. 2020, pp. 1541–1545. doi: 10.1190/segam2020-3428150.1
PDF Slides Code Link BibTeX

Mi Zhang, Ali Siahkoohi, and Felix J. Herrmann. Transfer learning in large-scale ocean bottom seismic wavefield reconstruction. In: SEG Technical Program Expanded Abstracts. Sept. 2020, pp. 1666–1670. doi:10.1190/segam2020-3427882.1.
PDF Slides Code Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, and Felix J. Herrmann. Weak deep priors for seismic imaging. In: SEG Technical Program Expanded Abstracts. Sept. 2020, pp. 2998–3002. doi: 10.1190/segam2020-3417568.1.
PDF Slides Code Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, and Felix J. Herrmann. Uncertainty quantification in imaging and automatic horizon tracking — a Bayesian deep-prior based approach. In: SEG Technical Program Expanded Abstracts. Sept. 2020, pp. 1636–1640. doi: 10.1190/segam2020-3417560.1.
PDF Slides Code Link BibTeX

Ali Siahkoohi, Gabrio Rizzuti, and Felix J. Herrmann. A deep-learning based Bayesian approach to seismic imaging and uncertainty quantification. In: 82nd EAGE Conference and Exhibition 2020. Jan. 2020.
PDF Slides Code Link BibTeX

Felix J. Herrmann, Ali Siahkoohi, and Gabrio Rizzuti. Learned imaging with constraints and uncertainty quantification. In: NeurIPS 2019 Deep Inverse Workshop. Dec. 2019.
PDF Slides Poster Link BibTeX

Ali Siahkoohi, Mathias Louboutin, and Felix J. Herrmann. Neural network augmented wave-equation simulation. In: Tech. rep. TR-CSE-2019-1, Georgia Institute of Technology. Sep. 2019.
PDF Code Link BibTeX

Ali Siahkoohi, Rajiv Kumar, and Felix J. Herrmann. Deep-learning based ocean bottom seismic wavefield recovery. In: SEG Technical Program Expanded Abstracts. Aug. 2019, pp. 2232–2237. doi: 10.1190/segam2019-3216632.1.
PDF Code-TensorFlow Code-PyTorch Slides Link BibTeX

Ali Siahkoohi, Dirk J. Verschuur, and Felix J. Herrmann. Surface-related multiple elimination with deep learning. In: SEG Technical Program Expanded Abstracts. Aug. 2019, pp. 4629–4634. doi: 10.1190/segam2019-3216723.1.
PDF Slides Link BibTeX

Ali Siahkoohi, Mathias Louboutin, and Felix J. Herrmann. The importance of transfer learning in seismic modeling and imaging. In: Geophysics 84.6. July 2019, pp. A47–A52. doi: 10.1190/geo2019-0056.1.
PDF Code Link BibTeX

Gabrio Rizzuti, Ali Siahkoohi, and Felix J. Herrmann. Learned iterative solvers for the Helmholtz equation. In: 81st EAGE Conference and Exhibition 2019. June 2019. doi: 10.3997/2214-4609.201901542.
PDF Slides Link BibTeX

Felix J. Herrmann, Ali Siahkoohi, and Mathias Louboutin. Machine Learning in Seismic Imaging — from Low-fidelity to High-fidelity. In: SIAM Conference on Computational Science and Engineering. (SIAM CSE). Mar. 2019.
Slides Link BibTeX

Ali Siahkoohi, Mathias Louboutin, Rajiv Kumar, and Felix J. Herrmann. Deep-convolutional neural networks in prestack seismic — two exploratory examples. In: SEG Technical Program Expanded Abstracts. Oct. 2018, pp. 2196–2200. doi: 10.1190/segam2018-2998599.1.
PDF Poster Link BibTeX

Felix J. Herrmann, Gerard J. Gorman, Jan Hückelheim, Keegan Lensink, Paul Kelly, Navjot Kukreja, Henryk Modzelewski, Michael Lange, Mathias Louboutin, Fabio Luporini, Ali Siahkoohi, and Philipp A. Witte. The power of abstraction in Computational Exploration Seismology. In: Smoky Mountains Computational Sciences and Engineering Conference. Aug. 2018.
Slides Link BibTeX

Ali Siahkoohi, Rajiv Kumar, and Felix J. Herrmann. Seismic Data Reconstruction with Generative Adversarial Networks. In: 80th EAGE Conference and Exhibition 2018. June 2018. doi: 10.3997/2214-4609.201801393.
PDF Slides Link BibTeX

Ali Siahkoohi and Ali Gholami. Sparsity Promoting Least Squares Migration for Laterally Inhomogeneous Media. In: 7th EAGE Saint Petersburg International Conference and Exhibition. Apr. 2016. doi: 10.3997/2214-4609.201600223.
PDF Link BibTeX

Mohmmad Sadegh Ebrahimi, Mohammad Hossein Daraei, Jamshid Rezaei, and Ali Siahkoohi. A Novel Utilization of Wireless Sensor Networks as Data Acquisition System in Smart Grids. In: Materials Science and Information Technology. Vol. 433. Advanced Materials Research. Trans Tech Publications, Jan. 2012, pp. 6725–6730. doi: 10.4028www.scientific.netAMR.433-440.6725.
PDF Link BibTeX

Amir Najafi, Ali Siahkoohi, and Mohammad B Shamsollahi. A content-based digital image watermarking algorithm robust against JPEG compression. In: 2011 IEEE International Symposium on Signal Processing and Information Technology (ISSPIT). IEEE. Feb. 2011, pp. 432–437. doi: 10.1109/ISSPIT.2011.6151601.
PDF Link BibTeX

Miscellaneous

“I wish you hadn't talked so much, it was distracting.”  —  Patrick Winston, How to Speak

“No one cares about the inside of your head.”  —  Larry McEnerney, The Craft of Writing Effectively