News Archive
2024
March 2024: Our preprint on a new multifidelity machine learning approach to learning from scarce data is available on arXiv. The work exploits the structure of linear regression problems to learn models from scientific and engineering data in a more robust way, leading to more accurate learned models when training data are scarce. This work is a collaboration between myself and PhD student Dayoung Kang with Anirban Chaudhuri and Vignesh Sella from UT Austin.
February 2024: I attended the SIAM Conference on Uncertainty Quantification held in Trieste, Italy from February 27 to March 1, where Peng Chen and I co-organized a minisymposium on Scalable algorithms for Bayesian inverse problems, and also presented joint work with PhD student Pavlos Stavrinides on balanced truncation for ensemble Kalman inversion.
January 2024: Welcome to two new undergraduate researchers, Holden Rohrer and Ben Zabriskie!
PhD student Tomoki Koike presented a paper at AIAA SciTech in session MDO-08: Metamodeling, Reduced Order Models, and Approximation Methods I. I chaired this session as well as MDO-09: Machine Learning and Optimization.
2023
December 2023: My proposal to Air Force Office of Scientific Research (AFOSR) Young Investigator Program (YIP) was selected for funding. The grant will support research on model reduction methods for inference problems. More info in the AFOSR press release.
November 2023: It was a privilege to give a talk at Virginia Tech on November 4 at the workshop honoring Christopher Beattie's 70th birthday.
October 2023: I am a co-PI on the ROME: Reduced Modeling from Extreme Data project, a DOE-funded multi-institution collaboration that will develop new model learning methods for Energy Earthshot applications. More information in the DOE press release.
I attended the Oakridge National Lab Core Universities AI workshop hosted at Georgia Tech, Oct 31-Nov 1.
I gave the virtual Oakridge National Lab AI Seminar on October 12.
I gave the Applied and Computational Mathematics Seminar at Georgia Tech on October 2.
September 2023: I have been awarded a Hans Fischer Fellowship by the Institute for Advanced Study at the Technical University of Munich (TUM). The fellowship will support a research collaboration with TUM researchers on reduced modeling for structural reliability analysis.
I was honored to be one of 81 early career engineers selected for the Grainger Foundation Frontiers of Engineering 2023 Symposium in Boulder, Colorado, September 10-13.
I attended ENUMATH 2023 in Lisbon, Portugal, September 4-8 and presented our work on balanced truncation and ensemble Kalman inversion.
August 2023: Welcome to new PhD students Dayoung Kang and Pavlos Stavrinides, and to new undergraduate researcher Jaffa Heryudono!
At ICIAM 2023 (Tokyo, August 20-25), I presented our work on the cost-accuracy trade-off in learning PDE operators with neural networks.
July 2023: I attended the US National Congress on Computational Mathematics (USNCCM) in Albuquerque, NM, July 23-27, 2023, where I presented our work on balanced truncation for Bayesian inference.
June 2023: I attended the workshop on Scientific Machine Learning at the Banff International Research Station, June 18-23.
Rising GT senior Han Zhang joined the group as a summer undergraduate research student. Han is an aerospace engineering major interested in data-driven modeling.
May 2023: ACE Group PhD student Tomoki and I both attended the conference on Nonlinear Model Reduction and Control hosted at Virginia Tech, May 22-26.
I visited UMass Amherst on May 5 to present our work on the cost-accuracy trade-off in learning PDE operators with neural networks in the reading seminar on mathematics of machine learning.
April 2023: I gave an invited plenary at the MASCOT-NUM 2023 conference in Le Croisic, France (April 3-6).
March 2023: I virtually presented our work on the cost-accuracy trade-off in learning PDE operators with neural networks in the Johns Hopkins University postdoc seminar in the Department of Applied Mathematics and Statistics on March 14.
February 2023: I presented our work on multifidelity global sensitivity analysis for the JW Space Telescope at SIAM CSE 23 in Amsterdam (Feb 26 - Mar 3). GT CSE Communications Officer Bryant Wine wrote a piece highlighting this work and other GT CSE contributions at the conference.
2022
December 2022: I gave an invited tutorial on operator learning using neural networks in the Remote Colloquium on Vortex Dominated Flows on Dec 9.
November 2022: I gave a virtual talk in the University of Waterloo Numerical Analysis and Scientific Computing seminar on Nov 29.
PhD student Tomoki Koike is the inaugural member of the Aerospace Computational Engineering (ACE) Group at Georgia Tech. Tomoki completed his BS in Aerospace Engineering at Purdue and his research interests are in model reduction and control. Welcome Tomoki!
September 2022: I gave an invited plenary at the Symposium on Inverse Problems in Potsdam, Germany. I also presented at the Model Reduction and Surrogate Modeling (MORE) Conference in Berlin.
Our paper on the cost-accuracy trade-off in operator learning with neural networks has been accepted for publication in the Journal of Machine Learning.
August 2022: I received a Best Presentation Award in the postdoctoral category from the IACM Female Researchers Chapter for my video presentation on Reduced operator inference for nonlinear PDEs at the virtual World Congress on Computational Mechanics (July 31--August 5).
July 2022: New paper with collaborators from NASA Goddard: Multifidelity uncertainty quantification and model validation of large-scale multidisciplinary systems. We propose a new multi-fidelity method for sensitivity analysis and apply the method to the JW Space Telescope thermal models. Our method accelerates the computation time from more than two months to just 2 days.
Our paper on reduced operator inference for nonlinear PDEs has appeared in the SIAM Journal of Scientific Computing (click here for the arXiv version).
I presented our paper on reduced operator inference for nonlinear PDEs at the hybrid SIAM Annual Meeting.
June 2022: I attended the workshop celebrating 30 years of Acta Numerica in Bedlewo, Poland, from June 26 to July 2.
I presented our work on the cost-accuracy trade-off in learning neural operators on June 21 at the US National Congress on Theoretical and Applied Mechanics in Austin, TX.
I received a 2021-2022 ASCIT Teaching Award from the Associated Students of Caltech. Nomination and selection for this award is run entirely by students.
I presented our work on Balanced Truncation for Bayesian Inference in the Oxford Computational Mathematics & Applications Seminar on June 2.
May 2022: My contributions toward advancing DEI at Caltech have been recognized with the departmental Gradient for Change Award as well as the division-wide New Horizons Award.
April 2022: I presented our work on Balanced truncation for Bayesian inference at SIAM UQ 2022, where I also co-organized with Sven Wang a minisymposium on Statistical and Computational Guarantees for Bayesian Inverse Problems.
March 2022: New manuscript on the cost-accuracy trade-off in operator learning with neural networks is out on arXiv. This work with Daniel Huang, Andrew Stuart, and Maarten de Hoop provides detailed numerical studies of the complexity question for neural net approximations of PDE-governed mappings between function spaces.
Our paper, Reduced operator inference for nonlinear partial differential equations, has been accepted for publication in the SIAM Journal on Scientific Computing.
February 2022: Our paper, Model reduction for linear dynamical systems via balancing for Bayesian inference, has been accepted for publication in the Journal of Scientific Computing. This work is a collaboration that grew out of the ICERM Semester Program on Model and Dimension Reduction.
Jan/Feb 2022: As chair of Caltech's Computational Mathematics + X Seminar in our winter term, I was delighted to host speakers Alexandria Volkening, Talea Mayo, Ron Buckmire, and Chad Topaz.
2021
December 2021: I presented our work on Balanced truncation for Bayesian inference at the Women in Inverse Problems virtual workshop hosted by Banff International Research Station Dec 5-10.
November 2021: Our manuscript, Model reduction for linear dynamical systems via balancing for Bayesian inference, is on arXiv. This work is a collaboration that grew out of the ICERM Semester Program on Model and Dimension Reduction.
October 2021: Attended the NextProf Nexus workshop at the University of Michigan Oct 4-7.
I presented our work on Balanced truncation for Bayesian inference at the UC San Diego Center for Control Systems and Dynamics Seminar on Oct 1.
September 2021: I presented our new function-space formulation of Operator Inference and Lift & Learn for learning reduced models for nonlinear PDEs at the hybrid MMLDT-CSET 2021 conference Sept 26-29.
July 2021: I gave a talk at SIAM AN on July 22 on Model reduction of linear dynamical systems via balancing for Bayesian inference.
April 2021: Invited to give a SCAN Seminar at Cornell on April 19.
March 2021: At SIAM CSE 21, I co-organized a mini-symposium on "Dimension reduction for Bayesian inverse problems" and also presented my thesis work. Recordings of all talks will be available to conference registrants until June 4 via the conference platform.
February 2021: New pre-print on "Reduced operator inference for nonlinear partial differential equations" with Ionut-Gabriel Farcas and Karen Willcox is up on arXiv. This work presents a new formulation for learning from data the operators of a reduced model that maps between Hilbert spaces, yielding speed-ups of 5-6 orders of magnitude for a 3D combustion simulation.
January 2021: Excited to be joining the Diversity, Equity, and Inclusion (DEI) Steering Committee in my new department.
1/1/21 - Happy New Year! January 1st is my start date for my new gig as von Karman Postdoctoral Instructor at Caltech. After nearly a decade at MIT, I'm looking forward to new opportunities and new challenges.
2020
December 2020: Very happy to have successfully defended my PhD thesis on December 7.
October 2020: I am co-organizing with Andrew Stuart a virtual mini-symposium on "Dimension reduction for Bayesian inverse problems" at SIAM CSE 21 in March. Looking forward to it.
August 2020: I attended the virtual workshop on Numerical Analysis for Data Science at SAMSI from August 26-27.
MIT held a Day of Dialogue on racism for August 5. Benjamin Zhang and I co-led a discussion session titled "Asian in America: Reflections on past and present".
July 2020: Awarded a 2020 SIAM Student Paper Prize for my work on multifidelity estimation of global sensitivity indices (read more at the SIAM News Blog). I also gave a presentation in the 1st Workshop on Scientific-Driven Deep Learning on July 1 (video available at the link).
June 2020: I gave a LANS Seminar on June 3 at Argonne National Lab.
May 2020: I attended in the Erwin Schrödinger Institute's virtual workshop on "Multilevel and multifidelity sampling methods in UQ for PDEs" May 4-5.
April 2020: I attended ICERM's virtual workshop on Computational Statistics and Data-Driven Models (April 20-24).
March 2020: I attended ICERM's virtual workshop on Algorithms for Dimension and Complexity Reduction (March 23-27).
February 2020: Our paper, "Lift & Learn: Physics-informed machine learning for large-scale nonlinear dynamical systems," has been accepted for publication the Physica D Special Issue on Machine Learning and Dynamical Systems. This is joint work with Boris Kramer, Benjamin Peherstorfer, and Karen Willcox, and can be downloaded from arXiv here.
Spring 2020: I spent the spring semester visiting ICERM for the semester program on "Model and dimension reduction in uncertain and dynamic systems".
2019
October 2019: Invited to present in the Boeing Applied Math Seminar on October 30.
September 2019: Invited to present in UMass Amhert's Applied Mathematics and Computation Seminar series on September 23.
July 2019: I attended the Women in Data Science & Mathematics workshop at ICERM July 29 - August 2.
June 2019: I attended the AIAA Aviation Forum in Dallas, TX and presented our paper, "Transform & Learn: A data-driven approach to model reduction." The paper can be downloaded here and code can be downloaded from GitHub.
May 2019: I attended the Women in Aerospace Symposium hosted at MIT May 28-29. I also visited and presented my research at Friedrich-Alexander-Universität Erlangen-Nürnberg and the Technical University of Munich the week of May 20th as a joint winner of the BGCE Student Paper Prize.
April 2019: I successfully defended my thesis proposal on April 8! I also participated in the Rising Stars in Computational and Data Sciences workshop on April 9-10 at UT Austin.
March 2019: I presented a poster as part of the MIT Center for Computational Engineering Symposium on Monday, March 18. I also gave an MIT SIAM Student Chapter Seminar on Thursday, March 14.
February 2019: I attended the SIAM CSE conference in Spokane, WA at the end of February and presented my doctoral thesis work on learning polynomial models for nonlinear PDEs in MS 301 on Model Reduction and Reduced-order Modeling of Dynamical Systems as well as in the BGCE Student Paper Prize finalist talks. I was joint winner of the BGCE Student Paper Prize competition, along with Zakia Zainib (SISSA).
January 2019: I attended the ICERM Workshop on Scientific Machine Learning at the end of January and presented a poster on my doctoral thesis work in learning polynomial models for nonlinear PDEs.