Upcoming Seminars

Learning neural operators accurately, efficiently, reliably, and in one shot

Posted on by

dr. lu

Speaker: Dr. Lu Lu – Yale University
Date: Sep 20, 2024; Time: 2:30 PM Location: PWEB 175

Abstract: As an emerging paradigm in scientific machine learning, deep neural operators pioneered by us can learn nonlinear operators of complex dynamic systems via neural networks. In this talk, I will present the deep operator network (DeepONet) to learn various operators that represent deterministic and stochastic differential equations. I will also present several extensions of DeepONet, such as DeepM&Mnet for multiphysics problems, DeepONet with proper orthogonal decomposition or Fourier decoder layers, MIONet for multiple-input operators, and multifidelity DeepONet. I will demonstrate the effectiveness of DeepONet and its extensions to diverse multiphysics and multiscale problems, such as bubble growth dynamics, high-speed boundary layers, electroconvection, hypersonics, geological carbon sequestration, full waveform inversion, and astrophysics. Deep learning models are usually limited to interpolation scenarios, and I will quantify the extrapolation complexity and develop a complete workflow to address the challenge of extrapolation for deep neural operators. Moreover, I will present the first operator learning method that only requires one PDE solution, i.e., one-shot learning, by introducing a new concept of local solution operator based on the principle of locality of PDEs.

Biographical Sketch: Dr. Lu Lu is an Assistant Professor in the Department of Statistics and Data Science at Yale University. Prior to joining Yale, he was an Assistant Professor in the Department of Chemical and Biomolecular Engineering at University of Pennsylvania from 2021 to 2023, and an Applied Mathematics Instructor in the Department of Mathematics at Massachusetts Institute of Technology from 2020 to 2021. He obtained his Ph.D. degree in Applied Mathematics at Brown University in 2020, master’s degrees in Engineering, Applied Mathematics, and Computer Science at Brown University, and bachelor’s degrees in Mechanical Engineering, Economics, and Computer Science at Tsinghua University in 2013. His current research interest lies in scientific machine learning, including theory, algorithms, software, and its applications to engineering, physical, and biological problems. His broad research interests focus on multiscale modeling and high performance computing for physical and biological systems. He has received the 2022 U.S. Department of Energy Early Career Award, and 2020 Joukowsky Family Foundation Outstanding Dissertation Award of Brown University. He is also an action editor of Journal of Machine Learning.

Morphology, optical properties & climate impact of soot nanoparticles

Posted on by

Abstract: Soot is a major air pollutant produced by incomplete combustion of hydrocarbon fuels. The contribution of soot to global warming is currently estimated with large uncertainty (partly) due to the fractal-like agglomerate structure of its constituent nanoparticles. Here, the dynamics of soot nanoparticles are investigated to advance our current understanding of particle formation during combustion. Discrete element modeling (DEM) enables the detailed description of the particle morphology (doi.org/10.1016/j.proci.2016.08.078) and optical properties (doi.org/10. 1016/j.proci.2018. 08.025) in population balance models and computational fluid dynamics (doi.org/10.1016/j.combustflame.2021.01.010). Power laws relating the optical properties of soot to its filamentary structure are derived by DEM (doi.org/10.1016/j.carbon.2017.06.004) to facilitate the accurate monitoring of soot emissions by aerosol (doi.org/10.1016/j.proci.2020. 07.055), laser (doi.org/10.1016/j.combustflame.2022.112025) diagnostics and fire detectors (doi.org/10.1016/j.powtec.2019.02.003). Most importantly, these relations enable the estimation of the soot direct radiative forcing accounting for its realistic agglomerate structure (doi.org/10.1021/acs.est.2c00428).

Biographical Sketch: Dr. Georgios Kelesidis is an Assistant Professor at Rutgers School of Public Health and Deputy Director of the Nanoscience and Advanced Materials Center of the Environmental and Occupational Health Sciences Institute at Rutgers University. Prior to this appointment, he was a Lecturer and Research Associate at the Department of Mechanical and Process Engineering of ETH Zürich, Switzerland. He received a Diploma in Chemical Engineering from the University of Patras, Greece with honors (top 3%), along with the Limmat Stiftung Award of Academic Excellence (2013). His subsequent MSc studies in Process Engineering at ETH Zürich were supported by a Particle Technology Laboratory Fellowship (2013-2015), while his MSc thesis earned the IBM research prize (2017) for computer modelling and simulations in chemistry, biology and material science. His 2019 PhD thesis on the morphology and optical properties of flame-made nanoparticles received the 2020 PhD Award from GAeF (German Association for Aerosol Research) and the ETH medal for Outstanding Doctoral Thesis (top 8 %). He received also the 1st Graduate Student Award on Carbon Nanomaterials at the 2019 AIChE Annual Meeting (Orlando, FL, USA), as well as Best Poster Awards at the European Aerosol Conference (EAC) in 2016 (Tours, France) and 2020 (Aachen, Germany), the 2019 ETH Conference on Combustion Generated Nanoparticles (Zürich, Switzerland) and the 2019 Fall Meeting of the Material Research Society (MRS). The societal impact of his PhD research was also highlighted by the Forbes Magazine by including him in the 2020 Forbes 30 under 30 Europe list for Science & Healthcare. He has co-authored 21 peer-reviewed articles so far, being the first author in 16 of them. He has organized technical sessions at MRS (2016), EAC (2019-2021), the 2020-2022 Annual Meetings of the American Association for Aerosol Research, the 11th International Aerosol Conference (2022) and the 9th World Congress on Particle Technology (2022). He has supervised so far 10 MSc and 7 BSc students. He is currently co-supervising 1 PhD student at ETH Zürich.