SciML & UQ Group

News

Papers, talks, awards, and people, since the group started in 2022. For shorter updates, follow Peng on LinkedIn.

2026

October 6, 2026
Our open-source library GeoSWE (Geophysical Shallow-Water Engine) is released: a GPU finite-volume solver for the 2D shallow-water equations, for flood modeling from a 3 m coastal county to the conterminous United States on one multi-GPU node. See the documentation.
October 2, 2026
Our paper LD-EnFF: Latent-Dynamics Ensemble Flow Filtering for Data Assimilation with Sparse Observations is submitted, with Ziyu Tian, Kaichen Shen, Wenbo Hao, Phillip Si, and Wei Zhu.
September 30, 2026
Our paper C-STRIDE: An Observation-Driven AI Digital Twin for Predicting Basin-Wide Flood Fields from Sparse Stream-Gauge Histories is submitted. Congratulations Yanjie, Phillip, and Yuan!
September 27, 2026
A collaborated paper Neural Scaling Laws of Transformer Operator Network is submitted, with Haoran Yan, Zhongjie Shi, Yuanzhe Xi, and Wenjing Liao.
September 24, 2026
Our paper LEVDA: Latent Ensemble Variational Data Assimilation via Differentiable Dynamics is accepted at NeurIPS 2026. Congratulations Phillip!
September 21, 2026
Our open-source library hIPPyMFEM is released: differentiable programming for infinite-dimensional Bayesian inverse problems, using MFEM on GPUs. See the documentation.
September 19, 2026
Our paper Toward AI-driven digital twins for metropolitan floods: A conditional latent dynamics network surrogate of the shallow water equations is published in the Journal of Hydrology, with code, data, and trained models at CLDNet. Congratulations Phillip, Yuan, and all!
September 14, 2026
Our paper Gaussian mixture Taylor approximations of risk measures constrained by PDEs with Gaussian random field inputs is accepted for publication in the SIAM/ASA Journal on Uncertainty Quantification. Congratulations DC and all!
September 7, 2026
Our paper Dimension reduction for derivative-informed operator learning: An analysis of approximation errors is published in the Journal of Machine Learning Research. Congratulations DC and all!
August 2, 2026
A collaborated paper PolyMicros: A foundation model for polycrystalline material structure is published in Acta Materialia, with Michael Buzzy, Andreas Robertson, and Surya Kalidindi.
July 22, 2026
Peng is selected to work on two projects of the Department of Energy's Genesis Mission, partnering with Oak Ridge National Laboratory and Argonne National Laboratory to improve predictions of water availability and flood risks.
July 17, 2026
Our paper Variationally correct operator learning: Reduced basis neural operator with a posteriori error estimation is published in Computer Methods in Applied Mechanics and Engineering. Congratulations Yuan!
July 6–10, 2026
Peng, Wenbo, and Yuan present two talks at the SIAM Annual Meeting (AN26) in Cleveland, on Gaussian mixture Taylor approximations for uncertainty quantification of control objectives and on variationally correct operator learning with a posteriori error estimation.
May 13, 2026
Our paper Toward AI-Driven Digital Twins for Metropolitan Floods: A Conditional Latent Dynamics Network Surrogate of the Shallow Water Equations is submitted. Congratulations Phillip, Yuan, and all!
March 22–25, 2026
Peng has 5 co-authored talks at the SIAM Conference on Uncertainty Quantification (UQ26) in Minneapolis: Kaichen on sequential Bayesian optimal experimental design via policy gradient reinforcement learning, Yuan on variationally stable reduced basis neural operators, Phillip on latent ensemble score filters for data assimilation with sparse observations, Ben on projected Stein variational inference using derivative-informed neural operators, and Peng on derivative-informed recurrent neural operators for sequential Bayesian optimal experimental design.
March 22, 2026
A collaborated paper Tucker Tensor Train Taylor Series is submitted, with Nick Alger, Blake Christierson, and Omar Ghattas.
March 4, 2026
Yanjie presents STRIDE for spatiotemporal reconstruction from sparse sensors at the HotCSE Seminar.
February 23, 2026
Our paper LEVDA: Latent Ensemble Variational Data Assimilation via Differentiable Dynamics is submitted. Congratulations Phillip!
February 5, 2026
Peng gives a seminar at Oak Ridge National Laboratory (Mathematics in Computation Section) on real-time and robust digital twins for learning and optimizing complex physical systems.
February 4, 2026
Our paper From Sparse Sensors to Continuous Fields: STRIDE for Spatiotemporal Reconstruction is submitted. Congratulations Yanjie!
January 26, 2026
Our paper LD-EnSF: Synergizing Latent Dynamics with Ensemble Score Filters for Fast Data Assimilation with Sparse Observations is accepted at ICLR 2026. Congratulations Pengpeng and Phillip!
January 21, 2026
Peng speaks at the Georgia Tech Machine Learning Seminar Series on scientific machine learning and uncertainty quantification for digital twins of physical systems.
January 9, 2026
Our paper Sequential Bayesian Optimal Experimental Design in Infinite Dimensions via Policy Gradient Reinforcement Learning is submitted. Congratulations Kaichen!

2025

December 24, 2025
Our paper Variationally correct operator learning: Reduced basis neural operator with a posteriori error estimation is submitted. Congratulations Yuan! This is a joint work with Prof. Wolfgang Dahmen at the University of South Carolina.
November 30–December 7, 2025
Our paper PDPO: Parametric Density Path Optimization is presented at NeurIPS 2025.
November 10–14, 2025
Peng co-organizes the IMSI workshop on reduced-order and surrogate modeling for digital twins.
September 19, 2025
Our paper PDPO: Parametric Density Path Optimization is accepted at NeurIPS 2025. Congratulations Sebastian!
August 8, 2025
Peng receives a grant from IDEaS + Cloud Hub RFP: Generative Artificial Intelligence (GenAI) for Science; our project deploys generative AI for real-time flood prediction. Thank you Cloud Hub and IDEaS!
July 20–24, 2025
Peng co-organizes a minisymposium and delivers an invited talk on uncertainty quantification and SciML at USNCCM18 in Chicago.
July 24, 2025
Our paper Efficient PDE-constrained optimization under high-dimensional uncertainty using derivative-informed neural operators is published in SIAM Journal on Scientific Computing. Congratulations DC and all!
July 1, 2025
Our paper Sequential infinite-dimensional Bayesian optimal experimental design with derivative-informed latent attention neural operator is published in the Journal of Computational Physics. Congratulations Jinwoo!
June 20, 2025
Congratulations to Benjamin Burns on receiving an NSF GRFP for research in machine learning.
June 10–12, 2025
Peng gives an invited talk at the DOE-ASCR Inverse Methods for Complex Systems under Uncertainty Workshop (Rockville, MD) on latent generative modeling for Bayesian data assimilation of complex dynamical systems.
June 6, 2025
Peng gives an invited talk at the ASA/IMS Spring Research Conference (Baruch College, NYC) on sequential infinite-dimensional Bayesian optimal experimental design with derivative-informed latent attention neural operators for optimal MRI tumor monitoring.
May 27, 2025
Sebastian Gutierrez Hernandez (co-advised by Prof. Haomin Zhou and Peng) submitted PDPO: Parametric Density Path Optimization.
May 22, 2025
A collaborated paper PolyMicros: Bootstrapping a Foundation Model for Polycrystalline Material Structure is submitted.
April 1, 2025
Our paper Accurate, scalable, and efficient Bayesian optimal experimental design with derivative-informed neural operators is published in Computer Methods in Applied Mechanics and Engineering. Congratulations Jinwoo!
March 31, 2025
Peng delivers an invited seminar at the Georgia Tech School of Mathematics on latent neural dynamics for fast data assimilation with sparse observations.
March 28, 2025
Peng speaks at the Clemson University School of Mathematical and Statistical Sciences on neural surrogates for fast and scalable Bayesian optimal experimental design.
January 24, 2025
Peng co-organizes the IMSI workshop on reduced-order and surrogate modeling for digital twins, scheduled for November 10–14, 2025.
January 22, 2025
Our paper Latent-EnSF: A Latent Ensemble Score Filter for High-Dimensional Data Assimilation with Sparse Observation Data is accepted at ICLR 2025. Congratulations Phillip!

2024

December 6, 2024
Congratulations to Irene Simó Muñoz for placing second in the ACM Student Research Competition at SC24; she studied at CSE in our group in Spring 2024.
December 3, 2024
Peng presents an invited talk at the Applied Mathematics Colloquium, Columbia University on neural surrogates for fast and scalable Bayesian optimal experimental design. Thanks Prof. Kui Ren for the invitation!
November 28, 2024
Our paper LD-EnSF: Synergizing Latent Dynamics with Ensemble Score Filters for Fast Data Assimilation with Sparse Observations is submitted. Congratulations Pengpeng and Phillip!
November 19, 2024
Peng presents an invited talk at the MATH4UQ Seminar, RWTH Aachen University on neural surrogates for fast and scalable Bayesian optimal experimental design.
November 18, 2024
Our paper Tempered multifidelity importance sampling for gravitational wave parameter estimation is published in Physical Review D, with Bassel Saleh, Aaron Zimmerman, and Omar Ghattas.
October 24, 2024
Peng collaborates on CRIS-HAZARD to collect essential flooding data from hurricanes Helene and Milton with University of South Florida and Georgia Tech researchers.
October 21–25, 2024
Peng co-organizes a minisymposium at SIAM MDS24 (Atlanta) on Data-Driven Scientific Machine Learning for Optimization of Complex Systems, with Prof. Raphaël Pestourie.
October 18, 2024
Our paper Optimal Neural Network Approximation of Wasserstein Gradient Direction via Convex Optimization is published in the SIAM Journal on Mathematics of Data Science.
October 13–15, 2024
Peng serves on the scientific committee for ROM+ML4LES+ at Emory University.
September 25, 2024
Our paper Derivative-enhanced Deep Operator Network is accepted at NeurIPS 2024. Congratulations Yuan and Nolan!
September 13, 2024
Our paper Sequential infinite-dimensional Bayesian optimal experimental design with derivative-informed latent attention neural operator is submitted. Congratulations Jinwoo!
August 29, 2024
Our paper Latent-EnSF: A Latent Ensemble Score Filter for High-Dimensional Data Assimilation with Sparse Observation Data is submitted. Congratulations Phillip!
August 19–21, 2024
Peng presents a talk at Workshop on Data-driven PDE-based inverse problem, in theory and practice, UW Madison on Neural Surrogates for Fast and Scalable Bayesian Optimal Experimental Design. Thanks Prof. Qin Li for the invitation.
August 15, 2024
Benjamin Burns, Wenbo Hao, and Yanjie Tong join us as PhD students. Welcome to all!
August 13, 2024
Our paper Gaussian mixture Taylor approximations of risk measures constrained by PDEs with Gaussian random field inputs is submitted. Congratulations Dingcheng and all!
August 5, 2024
Congratulations to Jinwoo for successfully defending his PhD Thesis on Efficient and Scalable Machine Learning Methods for Robust Bayesian Optimal Experimental Design!
July 21–26, 2024
Peng presents a talk at World Congress on Computational Mechanics 2024, Vancouver on Derivative-enhanced deep operator networks. Peng also co-organizes a minisymposium on UQ and SciML with Dr. Danial Faghihi et. al.
May 29, 2024
A collaborated paper Tempered Multifidelity Importance Sampling for Gravitational Wave Parameter Estimation is submitted. Congratulations Bassel and all!
April 6, 2024
Peng presents a plenary talk at Southeast Applied and Computational Math Student Workshop, 2024 on Derivative-informed Neural Operators for PDE-Constrained Optimization under Uncertainty. Yuan presents a poster on Derivative-enhanced Deep Operator Networks.
February 29, 2024
Our paper Derivative-enhanced deep operator network is submitted. Congratulations Yuan and Nolan!
February 28, 2024
Our paper Probabilistic Bayesian optimal experimental design using conditional normalizing flows is submitted. Congratulations Rafael Orozco! This is a joint work with Prof. Felix Herrmann at GT-CSE.
February 27–March 1, 2024
Peng has 8 co-authored talks at SIAM Conference on Uncertainty Quantification (UQ24). All PhD students present invited talks on their work, including Dingcheng on PDE-constrained optimization under uncertainty, Jinwoo on Bayesian optimal experimental design, Phillip on Semi-autoregressive energy flows, Sebastian on Control-DeepONet, and Yuan on derivative-enhanced DeepONet. Prof. Elizabeth Qian and Peng co-organize a minisymposium on Scalable Algorithms for Bayesian Inverse Problems.
February 24, 2024
Peng presents a plenary talk at Georgia Scientific Computing Symposium 2024 on Derivative-informed Neural Operators for PDE-Constrained Optimization under Uncertainty. Jinwoo, Phillip, Sebastian, and Yuan present four posters on Bayesian optimal experimental design, consistency model for data assimilation, Wasserstein gradient flows, and derivative-enhanced deep operator networks, respectively.
February 1, 2024
Peng speaks at Georgia Tech BBISS Seminar on Scientific Machine Learning for Predictive Digital Twins of Complex Systems.
January 19, 2024
Irene joins the group as a visiting Master student. Welcome Irene!
January 16, 2024
Peng co-organizes and presents at Georgia Tech Workshop on Foundation of Scientific AI for Optimization of Complex Systems.

2023

December 22, 2023
Our paper Accelerating Bayesian Optimal Experimental Design with Derivative-Informed Neural Operators is submitted. Congratulations Jinwoo!
December 21, 2023
Phillip and Sebastian receive the Student Travel Awards to present at SIAM UQ24. Congratulations Phillip and Sebastian!
December 19, 2023
Peng speaks at 2023 Workshop on Scientific Computing and Large Data at University of South Carolina on Derivative-Informed Neural Operator for PDE-Constrained Optimization under Uncertainty. Thanks Prof. Wuchen Li for the invitation!
December 4, 2023
Nathan receives the President’s Undergraduate Research Award (PURA) to conduct research on variational inference. Congratulations Nathan!
November 8, 2023
Peng speaks at 2023 BSRA/SRNL AI Workshop on Scientific Machine Learning for Prediction, Inference, Experimental Design, and Control of Complex Systems.
October 27, 2023
Our paper Derivative-Informed Neural Operator: An Efficient Framework for High-Dimensional Parametric Derivative Learning is published in Journal of Computational Physics. Congratulations Tom and all!
October 25, 2023
Peng speaks at Optimization in Oslo (OiO) Seminar on Derivative-Informed Neural Operator for PDE-Constrained Optimization under Uncertainty. Thanks Prof. Thomas M. Surowiec for the invitation!
October 16, 2023
Peng accepts an appointment as Brook Byers Institute for Sustainable Systems Faculty Fellow.
October 3, 2023
Peng receives an award of Thematic Events in AI from IDEaS to organize a workshop on Foundation of scientific AI (Artificial Intelligence) for Optimization of Complex Systems. Stay tuned.
September 21–22, 2023
Peng attends Oden Institute 50th Anniversary.
September 12, 2023
Peng speaks at Mechanical Engineering Seminar Series at University of Michigan. Thanks Prof. Xun Huan for the invitation!
September 1, 2023
Peng receives a Collaborative NSF Award (total $1.5M) on Variationally Stable Neural Networks for Simulation, Learning, and Experimental Design of Complex Physical Systems.
August 25–26, 2023
Peng speaks at Prof. Christoph Schwab 60th Birthday Workshop. Happy Birthday, Prof. Schwab!
August 21–25, 2023
Jinwoo and Peng speak at ICIAM 2023 Tokyo on Bayesian optimal experimental design and Bayesian inference, where Peng also co-organizes a minisymposium with Prof. Jinlong Wu on Scientific Machine Learning for Inverse Problems.
August 15, 2023
Phillip Si joins our group as a PhD student. Welcome Phillip!
August 10, 2023
A new version of our library SOUPy for stochastic optimization of complex systems under uncertainty is released. Congratulations and thanks to Dingcheng for leading this effort!
August 6, 2023
Peng presents a KDD'23 Tutorial on Uncertainty Quantification in Deep Learning. Thanks Lingkai for leading this tutorial!
July 17, 2023
Peng receives an NSF award as a CoPI on real-time flood prediction by data and model integration.
June 30, 2023
Peng speaks at DDPS Webinar, Lawrence Livermore National Laboratory on Derivative-informed neural operator. Thanks Dr. Siu Wun (Tony) Cheung for the invitation.
June 21, 2023
Jinwoo, Dingcheng, and Peng speak at Scientific Machine Learning, Banff International Research Station on Machine learning for Bayesian optimal experimental design, PDE-constrained optimization under uncertainty, and Bayesian inference problems. Thanks Dr. Thomas O'Leary-Roseberry for the invitation.
June 8, 2023
Our paper Bayesian model calibration for diblock copolymer thin film self-assembly using power spectrum of microscopy data is submitted. Congratulations Lianghao and all!
June 1, 2023
Peng presents at Mathematical and Scientific Machine Learning, ICERM on Derivative-informed neural operator for PDE-constrained optimization under uncertainty.
June 1, 2023
Peng speaks at SIAM Conference on Optimization (OP23) on Derivative-informed neural operator for PDE-constrained optimization under uncertainty.
May 31, 2023
Our paper Efficient PDE-Constrained optimization under high-dimensional uncertainty using derivative-informed neural operators is submitted. Congratulations Dingcheng and all!
April 28, 2023
Peng speaks at Isaac Newton Institute on Projected variational inference for high-dimensional Bayesian inverse problems.
April 21, 2023
Peng speaks at Auburn University on Scalable Bayesian optimal experimental design for efficient data acquisition. Thanks Prof. Yimin Zhong for the invitation.
March 27, 2023
Our paper Optimal design of chemoepitaxial guideposts for directed self-assembly of block copolymer systems using an inexact-Newton algorithm is accepted for publication in Journal of Computational Physics. Congratulations Dingcheng and all!
February 20, 2023
Peng speaks at GaTech Math on Scalable Bayesian optimal experimental design for efficient data acquisition. Thanks Prof. Wenjing Liao for the invitation.
February 2, 2023
Our paper Large-scale Bayesian optimal experimental design with derivative-informed projected neural network is accepted for publication in Journal of Scientific Computing. Congratulations Keyi and all!
January 27, 2023
Our paper A greedy sensor selection algorithm for hyperparameterized linear Bayesian inverse problems is submitted. Congratulations Nicole and all!
January 5, 2023
James Reineke joins the group as an undergraduate student working on SciML for optimal experimental design. Welcome James!

2022

December 7, 2022
Shikhar Jaiswal joins the group as a Master student working on SciML for stochastic optimization. Welcome Shikhar!
November 30, 2022
Peng joins the editorial board of Numerical Methods for Partial Differential Equations, an international journal that publishes the highest quality research in the rigorous analysis of novel techniques for the numerical solutions of partial differential equations (PDEs).
November 10, 2022
Peng speaks at Emory University in Atlanta. Thanks Prof. Mattias Chung for the invitation!
November 8, 2022
Nolan Bridge and Nathan Gong join the group as undergraduate students working on UQ for SciML. Welcome Nolan and Nathan!
November 4–6, 2022
Jinwoo attends and speaks at 5th Annual Meeting of the SIAM Texas-Louisiana Section in Houston.
October 31, 2022
Brian Zhang joins the group as an undergraduate student working on UQ and Bayesian inference. Welcome Brian!
October 21, 2022
Our paper A fast and scalable computational framework for large-scale and high-dimensional Bayesian optimal experimental design. is accepted for publication in SIAM / ASA Journal on Uncertainty Quantification. Congratulations Keyi and all!
September 26–October 2, 2022
Peng attends and speaks at SIAM Conference on Mathematics of Data Science (MDS22) in San Diego. With Thomas O'Leary-Roseberry and Omar Ghattas, we organize a minisymposium on "Advances in Fast and Scalable Bayesian Inference".
September 27, 2022
Peng speaks at The University of Iowa in Iowa City. Thanks Prof. Xueyu Zhu for the invitation!
September 9, 2022
Our paper "An offline--online decomposition method for efficient linear Bayesian goal-oriented optimal experimental design: Application to optimal sensor placement" is accepted for publication in SIAM Journal on Scientific Computing. Congratulations Keyi and all!
September 6, 2022
Our paper "Optimal design of photonic nanojets under uncertainty" is submitted. Congratulations Amal and all!
August 22, 2022
Yuan Qiu joins the group as a Master student working on SciML and neural operators. Welcome Yuan!
August 18, 2022
Yulong Liang joins the group as a Master student working on UQ for nuclear fusion. Welcome Yulong!
August 13, 2022
Jinwoo Go joins the group as a PhD student working on SciML for optimal experimental design. Welcome Jinwoo!
August 2, 2022
Our paper "Optimal design of chemoepitaxial guideposts for directed self-assembly of block copolymer systems using an inexact-Newton algorithm" is submitted. Congratulations Dingcheng and all!
August 1, 2022
Peng joins the School of CSE at Georgia Tech as an Assistant Professor.