Saviz Mowlavi

- Phone: 617-621-7588
- Email:
-
Position:
Research / Technical Staff
Visiting Research Scientist -
Education:
Ph.D., Massachusetts Institute of Technology, 2022 -
Research Areas:
Saviz's Quick Links
-
Biography
Saviz's research is motivated by the integration of machine learning and physical laws for solving high-dimensional inverse problems in the realm of fluid mechanics, solid mechanics, heat transfer and beyond. During his PhD, he developed computational methods for various physical problems such as detecting hidden voids or inclusions in elastic bodies, identifying coherent structures in ensembles of moving particles, or controlling unstable fluid flows.
-
Recent News & Events
-
TALK [MERL Seminar Series 2023] Prof. Shaowu Pan presents talk titled Neural Implicit Flow Date & Time: Wednesday, March 1, 2023; 1:00 PM
Speaker: Shaowu Pan, Rensselaer Polytechnic Institute
MERL Host: Saviz Mowlavi
Research Areas: Computational Sensing, Data Analytics, Machine LearningAbstractHigh-dimensional spatio-temporal dynamics can often be encoded in a low-dimensional subspace. Engineering applications for modeling, characterization, design, and control of such large-scale systems often rely on dimensionality reduction to make solutions computationally tractable in real-time. Common existing paradigms for dimensionality reduction include linear methods, such as the singular value decomposition (SVD), and nonlinear methods, such as variants of convolutional autoencoders (CAE). However, these encoding techniques lack the ability to efficiently represent the complexity associated with spatio-temporal data, which often requires variable geometry, non-uniform grid resolution, adaptive meshing, and/or parametric dependencies. To resolve these practical engineering challenges, we propose a general framework called Neural Implicit Flow (NIF) that enables a mesh-agnostic, low-rank representation of large-scale, parametric, spatial-temporal data. NIF consists of two modified multilayer perceptrons (MLPs): (i) ShapeNet, which isolates and represents the spatial complexity, and (ii) ParameterNet, which accounts for any other input complexity, including parametric dependencies, time, and sensor measurements. We demonstrate the utility of NIF for parametric surrogate modeling, enabling the interpretable representation and compression of complex spatio-temporal dynamics, efficient many-spatial-query tasks, and improved generalization performance for sparse reconstruction.
-
NEWS MERL researchers presenting workshop papers at NeurIPS 2022 Date: December 2, 2022 - December 8, 2022
MERL Contacts: Matthew Brand; Toshiaki Koike-Akino; Jing Liu; Saviz Mowlavi; Kieran Parsons; Ye Wang
Research Areas: Artificial Intelligence, Control, Dynamical Systems, Machine Learning, Signal ProcessingBrief- In addition to 5 papers in recent news (https://www.merl.com/news/news-20221129-1450), MERL researchers presented 2 papers at the NeurIPS Conference Workshop, which was held Dec. 2-8. NeurIPS is one of the most prestigious and competitive international conferences in machine learning.
- “Optimal control of PDEs using physics-informed neural networks” by Saviz Mowlavi and Saleh Nabi
Physics-informed neural networks (PINNs) have recently become a popular method for solving forward and inverse problems governed by partial differential equations (PDEs). By incorporating the residual of the PDE into the loss function of a neural network-based surrogate model for the unknown state, PINNs can seamlessly blend measurement data with physical constraints. Here, we extend this framework to PDE-constrained optimal control problems, for which the governing PDE is fully known and the goal is to find a control variable that minimizes a desired cost objective. We validate the performance of the PINN framework by comparing it to state-of-the-art adjoint-based optimization, which performs gradient descent on the discretized control variable while satisfying the discretized PDE.
- “Learning with noisy labels using low-dimensional model trajectory” by Vasu Singla, Shuchin Aeron, Toshiaki Koike-Akino, Matthew E. Brand, Kieran Parsons, Ye Wang
Noisy annotations in real-world datasets pose a challenge for training deep neural networks (DNNs), detrimentally impacting generalization performance as incorrect labels may be memorized. In this work, we probe the observations that early stopping and low-dimensional subspace learning can help address this issue. First, we show that a prior method is sensitive to the early stopping hyper-parameter. Second, we investigate the effectiveness of PCA, for approximating the optimization trajectory under noisy label information. We propose to estimate the low-rank subspace through robust and structured variants of PCA, namely Robust PCA, and Sparse PCA. We find that the subspace estimated through these variants can be less sensitive to early stopping, and can outperform PCA to achieve better test error when trained on noisy labels.
- In addition, new MERL researcher, Jing Liu, also presented a paper entitled “CoPur: Certifiably Robust Collaborative Inference via Feature Purification" based on his previous work before joining MERL. His paper was elected as a spotlight paper to be highlighted in lightening talks and featured paper panel.
- In addition to 5 papers in recent news (https://www.merl.com/news/news-20221129-1450), MERL researchers presented 2 papers at the NeurIPS Conference Workshop, which was held Dec. 2-8. NeurIPS is one of the most prestigious and competitive international conferences in machine learning.
-
-
MERL Publications
- "Reinforcement Learning-based Estimation for Partial Differential Equations", SIAM Conference on Applications of Dynamical Systems, May 2023.BibTeX TR2023-066 PDF
- @inproceedings{Mowlavi2023may,
- author = {Mowlavi, Saviz and Benosman, Mouhacine and Nabi, Saleh},
- title = {Reinforcement Learning-based Estimation for Partial Differential Equations},
- booktitle = {SIAM Conference on Applications of Dynamical Systems},
- year = 2023,
- month = may,
- url = {https://www.merl.com/publications/TR2023-066}
- }
, - "Optimal Control of PDEs Using Physics-Informed Neural Networks", Advances in Neural Information Processing Systems (NeurIPS) workshop, December 2022.BibTeX TR2022-163 PDF
- @inproceedings{Mowlavi2022dec,
- author = {Mowlavi, Saviz and Nabi, Saleh},
- title = {Optimal Control of PDEs Using Physics-Informed Neural Networks},
- booktitle = {Advances in Neural Information Processing Systems (NeurIPS) workshop},
- year = 2022,
- month = dec,
- url = {https://www.merl.com/publications/TR2022-163}
- }
, - "Optimal control of PDEs using physics-informed neural networks", Journal of Computational Physics, DOI: j.jcp.2022.111731, Vol. 473, pp. 111731, October 2022.BibTeX TR2022-143 PDF
- @article{Mowlavi2022oct,
- author = {Mowlavi, Saviz and Nabi, Saleh},
- title = {Optimal control of PDEs using physics-informed neural networks},
- journal = {Journal of Computational Physics},
- year = 2022,
- volume = 473,
- pages = 111731,
- month = oct,
- doi = {j.jcp.2022.111731},
- url = {https://www.merl.com/publications/TR2022-143}
- }
, - "Reinforcement Learning State Estimation for High-Dimensional Nonlinear Systems", International Conference on Learning Representations (ICLR) Workshop, April 2022.BibTeX TR2022-042 PDF
- @inproceedings{Mowlavi2022apr,
- author = {Mowlavi, Saviz and Benosman, Mouhacine and Nabi, Saleh},
- title = {Reinforcement Learning State Estimation for High-Dimensional Nonlinear Systems},
- booktitle = {International Conference on Learning Representations (ICLR) Workshop},
- year = 2022,
- month = apr,
- url = {https://www.merl.com/publications/TR2022-042}
- }
,
- "Reinforcement Learning-based Estimation for Partial Differential Equations", SIAM Conference on Applications of Dynamical Systems, May 2023.
-
Other Publications
- "Topology optimization with physics-informed neural networks: application to noninvasive detection of hidden geometries", arXiv preprint arXiv:2303.09280, 2023.BibTeX
- @Article{mowlavi2023topology,
- author = {Mowlavi, Saviz and Kamrin, Ken},
- title = {Topology optimization with physics-informed neural networks: application to noninvasive detection of hidden geometries},
- journal = {arXiv preprint arXiv:2303.09280},
- year = 2023
- }
, - "Detecting Lagrangian coherent structures from sparse and noisy trajectory data", Journal of Fluid Mechanics, Vol. 948, pp. A4, 2022.BibTeX
- @Article{mowlavi2022detecting,
- author = {Mowlavi, Saviz and Serra, Mattia and Maiorino, Enrico and Mahadevan, L},
- title = {Detecting Lagrangian coherent structures from sparse and noisy trajectory data},
- journal = {Journal of Fluid Mechanics},
- year = 2022,
- volume = 948,
- pages = {A4},
- publisher = {Cambridge University Press}
- }
, - "Contact model for elastically anisotropic bodies and efficient implementation into the discrete element method", Granular Matter, Vol. 23, No. 2, pp. 1-29, 2021.BibTeX
- @Article{mowlavi2021contact,
- author = {Mowlavi, Saviz and Kamrin, Ken},
- title = {Contact model for elastically anisotropic bodies and efficient implementation into the discrete element method},
- journal = {Granular Matter},
- year = 2021,
- volume = 23,
- number = 2,
- pages = {1--29},
- publisher = {Springer}
- }
, - "Interplay between hysteresis and nonlocality during onset and arrest of flow in granular materials", Soft Matter, Vol. 17, No. 31, pp. 7359-7375, 2021.BibTeX
- @Article{mowlavi2021interplay,
- author = {Mowlavi, Saviz and Kamrin, Ken},
- title = {Interplay between hysteresis and nonlocality during onset and arrest of flow in granular materials},
- journal = {Soft Matter},
- year = 2021,
- volume = 17,
- number = 31,
- pages = {7359--7375},
- publisher = {Royal Society of Chemistry}
- }
, - "Reduced model for capillary breakup with thermal gradients: Predictions and computational validation", Physics of Fluids, Vol. 33, No. 12, pp. 122003, 2021.BibTeX
- @Article{shukla2021reduced,
- author = {Shukla, Isha and Wang, Fan and Mowlavi, Saviz and Guyomard, Amy and Liang, Xiangdong and Johnson, Steven G and Nave, J-C},
- title = {Reduced model for capillary breakup with thermal gradients: Predictions and computational validation},
- journal = {Physics of Fluids},
- year = 2021,
- volume = 33,
- number = 12,
- pages = 122003,
- publisher = {AIP Publishing LLC}
- }
, - "Control of linear instabilities by dynamically consistent order reduction on optimally time-dependent modes", Nonlinear Dynamics, Vol. 95, No. 4, pp. 2745-2764, 2019.BibTeX
- @Article{blanchard2019control,
- author = {Blanchard, Antoine and Mowlavi, Saviz and Sapsis, Themistoklis P},
- title = {Control of linear instabilities by dynamically consistent order reduction on optimally time-dependent modes},
- journal = {Nonlinear Dynamics},
- year = 2019,
- volume = 95,
- number = 4,
- pages = {2745--2764},
- publisher = {Springer}
- }
, - "Particle size selection in capillary instability of locally heated coaxial fiber", Physical Review Fluids, Vol. 4, No. 6, pp. 064003, 2019.BibTeX
- @Article{mowlavi2019particle,
- author = {Mowlavi, Saviz and Shukla, Isha and Brun, P-T and Gallaire, Fran{\c{c}}ois},
- title = {Particle size selection in capillary instability of locally heated coaxial fiber},
- journal = {Physical Review Fluids},
- year = 2019,
- volume = 4,
- number = 6,
- pages = 064003,
- publisher = {APS}
- }
, - "Absolute/convective secondary instabilities and the role of confinement in free shear layers", Physical Review Fluids, Vol. 3, No. 5, pp. 053901, 2018.BibTeX
- @Article{arratia2018absolute,
- author = {Arratia, Crist{\'o}bal and Mowlavi, Saviz and Gallaire, Fran{\c{c}}ois},
- title = {Absolute/convective secondary instabilities and the role of confinement in free shear layers},
- journal = {Physical Review Fluids},
- year = 2018,
- volume = 3,
- number = 5,
- pages = 053901,
- publisher = {APS}
- }
, - "Model order reduction for stochastic dynamical systems with continuous symmetries", SIAM Journal on Scientific Computing, Vol. 40, No. 3, pp. A1669-A1695, 2018.BibTeX
- @Article{mowlavi2018model,
- author = {Mowlavi, Saviz and Sapsis, Themistoklis P},
- title = {Model order reduction for stochastic dynamical systems with continuous symmetries},
- journal = {SIAM Journal on Scientific Computing},
- year = 2018,
- volume = 40,
- number = 3,
- pages = {A1669--A1695},
- publisher = {SIAM}
- }
, - "Spatio-temporal stability of the Kármán vortex street and the effect of confinement", Journal of Fluid Mechanics, Vol. 795, pp. 187-209, 2016.BibTeX
- @Article{mowlavi2016spatio,
- author = {Mowlavi, Saviz and Arratia, Crist{\'o}bal and Gallaire, Fran{\c{c}}ois},
- title = {Spatio-temporal stability of the Kármán vortex street and the effect of confinement},
- journal = {Journal of Fluid Mechanics},
- year = 2016,
- volume = 795,
- pages = {187--209},
- publisher = {Cambridge University Press}
- }
, - "In vivo observations and in vitro experiments on the oral phase of swallowing of Newtonian and shear-thinning liquids", Journal of Biomechanics, Vol. 49, No. 16, pp. 3788-3795, 2016.BibTeX
- @Article{mowlavi2016vivo,
- author = {Mowlavi, S and Engmann, J and Burbidge, A and Lloyd, R and Hayoun, P and Le Reverend, B and Ramaioli, M},
- title = {In vivo observations and in vitro experiments on the oral phase of swallowing of Newtonian and shear-thinning liquids},
- journal = {Journal of Biomechanics},
- year = 2016,
- volume = 49,
- number = 16,
- pages = {3788--3795},
- publisher = {Elsevier}
- }
, - "A model experiment to understand the oral phase of swallowing of Newtonian liquids", Journal of biomechanics, Vol. 48, No. 14, pp. 3922-3928, 2015.BibTeX
- @Article{hayoun2015model,
- author = {Hayoun, P and Engmann, J and Mowlavi, S and Le Reverend, B and Burbidge, A and Ramaioli, M},
- title = {A model experiment to understand the oral phase of swallowing of Newtonian liquids},
- journal = {Journal of biomechanics},
- year = 2015,
- volume = 48,
- number = 14,
- pages = {3922--3928},
- publisher = {Elsevier}
- }
,
- "Topology optimization with physics-informed neural networks: application to noninvasive detection of hidden geometries", arXiv preprint arXiv:2303.09280, 2023.