Gradient-Free Generation for Hard-Constrained Systems
Cheng, C., Han, B., Maddix, D.C., Ansari, A.F., Stuart, A., Mahoney, M.W., Wang, Y., (2024). "Gradient-Free Generation for Hard-Constrained Systems." Under Review.
CV
Education
- Ph.D. in Computational and Mathematical Engineering, Stanford University, 2018
- M.S. in Computational and Mathematical Engineering, Stanford University, 2015
- B.A. in Applied Mathematics, University of California, Berkeley, 2012
Work Experience
- March 2021 - Present: Senior Applied Scientist, AWS AI Labs, Amazon Web Services (AWS)
- I am leading the DeepEarth research team to develop novel physics-constrained machine learning solutions for computational fluid dynamics (CFD) and earth science applications.
- September 2018 - March 2021: Applied Scientist II, AWS AI Labs, Amazon Web Services (AWS)
- I worked on developing, implementing and researching novel probabilistic deep learning models for time series forecasting in the open-source package GluonTS and leading the productionalization effort of these algorithms for Amazon Forecast.
- March 2016 - June 2018: Research Associate, Stanford University
- In my PhD research, I worked with Professor Gerritsen in developing accurate and stable methods, as well as computationally efficient algorithms, for modeling flow through porous media. These methods are designed to accurately represent the underlying physics, and remove numerical artifacts associated with current state-of-the-art finite volume discretizations.
- Advisor: Professor Margot Gerritsen
- June 2017 - September 2017: Extreme-Scale Data Analytics Intern
- I developed physics-constrained (structure preserving) reduced order models (ROMs) for computationally expensive high fidelity models in computational fluid dynamics.
- June 2015 - September 2015: CUDA Software Engineer Intern, NVIDIA
- I developed efficient parallel algorithms for sparse matrix vector multiplication (spmv) on the GPU for CUDA toolkit libraries, including cuSPARSE and nvGRAPH. I also worked on problems in graph analytics, including utilizing the spmv to accelerate the pagerank, maximum flow and single source shortest path algorithms.
- May 2012 - June 2015: Computational Researcher, Lawrence Berkeley National Laboratory (LBL)
- I worked in the Computational Research Division under Professor James Sethian, applying multiphase flow algorithms using the Voronoi Implicit Interface Method (VIIM) to shape optimization problems.
Publications
DrivAerML: High-Fidelity Computational Fluid Dynamics Dataset for Road-Car External Aerodynamics
Ashton, N., Mockett, C., Fuchs, M., Fliessbach, L., Hetmann, H., Knacke, T., Schönwald, N., Skaperdas, V., Fotiadis, G., Walle, A., Hupertz, B. Maddix, D.C., Yu, P., (2024). "DrivAerML: High-Fidelity Computational Fluid Dynamics Dataset For Road-Car External Aerodynamics." Technical Report, Preprint arXiv:2408.11969, Under Review.
WindsorML–High-Fidelity Computational Fluid Dynamics Dataset For Automotive Aerodynamics
Ashton, N., Angel, J.B., Ghate, A.S., Kenway, G.K.W., Wong, M.L., Kiris, C., Walle, A., Maddix, D.C., Page, G., (2024). "WindsorML: High-Fidelity Computational Fluid Dynamics Dataset For Automotive Aerodynamics." Technical Report, Preprint arXiv:2407.19320, NeurIPS Datasets and Benchmarks Track, Accepted for Publication.
Comparing and Contrasting Deep Learning Weather Prediction Backbones on Navier-Stokes and Atmospheric Dynamics
Karlbauer, M., Maddix, D.C., Ansari, A.F., Han, B., Gupta, G., Wang, Y., Stuart, A., Mahoney, M.W., (2024). "Comparing and Contrasting Deep Learning Weather Prediction Backbones on Navier-Stokes and Atmospheric Dynamics." Technical Report, Preprint arXiv:2407.14129, Under Review.
Using Uncertainty Quantification to Characterize and Improve Out-of-Domain Learning for PDEs
Mouli, S.C., Maddix, D.C., Alizadeh, S., Gupta, G., Wang, Y., Stuart, A., Mahoney, M.W. (2024). "Using Uncertainty Quantification to Characterize and Improve Out-of-Domain Learning for PDEs." Proceedings of the 41st International Conference on Machine Learning (ICML), PMLR. 235.
Transferring Knowledge from Large Foundation Models to Small Downstream Tasks
Qiu, S., Han, B, Maddix, D.C., Zhang, S., Wang, Y., Wilson, A.G. (2024). "Transferring Knowledge from Large Foundation Models to Small Downstream Tasks." Proceedings of the 41st International Conference on Machine Learning (ICML), PMLR. 235.
Chronos: Learning the Language of Time Series
Ansari, F.A., Stella, L., Turkmen, C., Zhang, X., Mercado, P., Shen, H., Shchur, O., Rangapuram, S.S., Arango, S.A., Kapoor, S., Zschiegner, J., Maddix, D.C., et al. (2024). "Chronos: Learning the Language of Time Series." Transactions on Machine Learning Research (10/2024).
Machine Learning for Road Vehicle Aerodynamics Simulation
Ananthan, V., Ashton, N., Chadwick, N., Lizarraga, M., Maddix, D.C., et al. (2024). "Machine Learning for Road Vehicle Aerodynamics Simulation." Society of Automotive Engineers (SAE) Technical Paper.
Learning Physical Models that Can Respect Conservation Laws
Hansen, D.*, Maddix, D.C.*, Alizadeh, S., Gupta, G., Mahoney, M.W. (2024). "Learning Physical Models that Can Respect Conservation Laws." Physica D: Nonlinear Phenomena, 457 (133952), (*Equal contributions).
PreDiff: Precipitation Nowcasting with Latent Diffusion Models
Gao, Z., Shi, X., Han, B., Wang, H., Jin, X., Maddix, D.C., Zhu, Y., Li, M., Wang, Y. (2023). "PreDiff: Precipitation Nowcasting with Latent Diffusion Models." Proceedings of the 37th Conference of Neural Information Processing Systems (NeurIPS).
Learning Physical Models that Can Respect Conservation Laws
Hansen, D., Maddix, D.C., Alizadeh, S., Gupta, G., Mahoney, M.W. (2023). "Learning Physical Models that Can Respect Conservation Laws." Proceedings of the 40th International Conference on Machine Learning (ICML), PMLR. 202:12469-12510.
Theoretical Guarantees of Learning Ensembling Strategies with Applications to Time Series Forecasting
Hasson, H., Maddix, D.C., Wang, Y., Park, Y., Gupta, G., (2023). "Theoretical Guarantees of Learning Ensembling Strategies with Applications to Time Series Forecasting." Proceedings of the 40th International Conference on Machine Learning (ICML), PMLR. 202:12616-12632.
Guiding Continuous Operator Learning through Physics-based boundary constraints
Saad, N.*, Gupta, G.*, Alizadeh, S., Maddix, D.C. (2023). "Guiding Continuous Operator Learning through Physics-based boundary constraints." Proceedings of the International Conference on Learning Representations (ICLR).
Deep Learning For Time Series Forecasting: Tutorial and Literature Survey
Benidis, K., Rangapuram, S., Flunkert V., Wang, Y., Maddix, D.C., Türkmen, C., Gasthaus, J., Bohlke-Schneider, M., Salinas, D., Stella, L., Aubet, FX., Callot, L, Januschowski, T. (2022). "Deep Learning for Time Series Forecasting: Tutorial and Literature Survey." ACM Computing Surveys. 55(6):1-36.
Domain Adaptation for Time Series Forecasting via Attention Sharing
Jin, X., Park, Y., Maddix, D.C., Wang, H., Wang, Y. (2022). "Domain Adaptation for Time Series Forecasting via Attention Sharing." Proceedings of the 39th International Conference on Machine Learning (ICML), PMLR. 162:10280-10297.
Learning Quantile Functions without Quantile Crossing for Distribution-free Time Series Forecasting
Park, Y., Maddix, D.C., Aubet, FX., Kan, K., Gasthaus, J., Wang, Y. (2022). "Learning Quantile Functions without Quantile Crossing for Distribution-free Time Series Forecasting." Proceedings of the 25th International Conference on Artificial Intelligence and Statistics (AISTATS), PMLR. 151:8127-8150.
Bridging Physics-based and Data-driven modeling for Learning Dynamical Systems
Wang, R., Maddix, D.C., Faloutsos, C., Wang, Y., Yu, R. (2021). "Bridging Physics-based and Data-driven modeling for Learning Dynamical Systems." Proceedings of the 3rd Conference on Learning for Dynamics and Control (L4DC), PMLR. 144:385-398.
GluonTS: Probabilistic and Neural Time Series Modeling in Python
Alexandrov, A., Benidis, K., Bohlke-Schneider, M., Flunkert, V., Gasthaus, J., Januschowski, T., Maddix, D.C., Rangapuram, S., Salinas, D., Schulz, J., Stella, L., Türkmen, A.C., Wang, Y. (2020). "GluonTS: Probabilistic and Neural Time Series Modeling in Python." Journal of Machine Learning Research (JMLR). 21(116):1-6.
Deep Factors for Forecasting
Wang, Y., Smola, A., Maddix, D.C., Gasthaus, J., Foster, D. (2019). "Deep Factors for Forecasting." Proceedings of the 39th International Conference on Machine Learning (ICML), PMLR. 97:6607-6617.
Numerical Artifacts in the discontinuous Generalized Porous Medium Equation: How to avoid spurious temporal oscillations
Maddix, D.C., Sampaio, L., Gerritsen, M. (2018). "Numerical Artifacts in the discontinuous Generalized Porous Medium Equation: How to avoid spurious temporal oscillations." Journal of Computational Physics. 368:277-298.
Numerical Artifacts in the Generalized Porous Medium Equation: Why harmonic averaging itself is not to blame
Maddix, D.C., Sampaio, L., Gerritsen, M. (2018). "Numerical Artifacts in the Generalized Porous Medium Equation: Why harmonic averaging itself is not to blame." Journal of Computational Physics. 361:280-298.
Advanced Fluid Reduced Order Models for Compressible Flow
Tezaur, I.K., Fike, J., Carlberg, K., Barone, M., Maddix, D.C., Mussoni, E., Balajewicz, M. (2017). "Advanced Fluid Reduced Order Models for Compressible Flow." Sandia National Laboratories Report, Sand No. 2017-10335.
Talks
Intersection of Foundation Models and Scientific Computing
Talk at Foundation Models for Science: Progress, Opportunities, and Challenges, NeurIPS 2024 Workshop, Vancouver, Canada
ICME’s 20th Anniversary Celebration Event
Talk at Institute of Computational and Mathematical Engineering (ICME), Stanford University
UC Berkeley Math Career Panel
Talk at UC Berkeley Mathematics Department, University of California, Berkeley
Physics-constrained Machine Learning for Scientific Computing
Talk at Ansys Monthly Seminar, Paris, France
Advances in Scientific Machine Learning (SciML) in Industry
Talk at ICERM Workshop on The Industrialization of SciML, Brown University
Physics-constrained Machine Learning for Scientific Computing
Talk at CS294 Graduate Course on Physics Inspired Deep Learning, University of California, Berkeley
Physics-constrained Machine Learning for Earth and Sustainability Science
Talk at 2nd AI for Good Webinar Series: AI for Earth and Sustainability Science, Virtual
Women in Science at Amazon: A Conversation with our Amazonians Panelists
Talk at Amazon Science, Virtual
Physics-constrained Machine Learning for Scientific Computing
Talk at Machine Learning and Dynamical Systems Seminar, Alan Turing Institute, London, England
Mathematics in Science
Talk at Daves Avenue Elementary School Science Fair, Los Gatos, CA
The Intersection of Computational Math and Industry
Talk at Computational Math in Industry and Beyond Seminar Series (CME 500), Stanford, CA
Neural Time Series Models with GluonTS
Talk at International Conference on Machine Learning (ICML) Time Series Workshop, Long Beach, CA
Temporal oscillations in the porous medium equation: why harmonic averaging itself is not to blame
Talk at SIAM Conference on Mathematical and Computational Issues in the Geosciences 2017, Erlangen, Germany
ICME Xpo Research Symposium
Talk at Institute for Computational and Mathematical Engineering, Stanford University, Stanford, CA
Teaching
Awards and Honors
- December 2020: Best Paper Award, Machine Learning in Public Health NeurIPS 2020 Workshop
- AutoODE: Bridging Physics-based and Data-driven modeling for COVID-19 Forecasting
- May 2018: ICME Outstanding Leadership Award
- May 2018: Senior Teaching Fellow
- May 2016: 2016 ICME Xpo Poster Presentation Award
- Sparse Matrix Vector Multiplication Using the Merge Path (NVIDIA Internship project)
- 2013-2016: 2013 National Science Foundation (NSF) Graduate Research Fellowship Program (GRFP)
- March 2013: 2012 – 2013 Outstanding Graduate Student Instructor Award, University of California, Berkeley
- May 2012: Percy Lionel Davis Award for Excellence in Mathematics
- Spring 2011: Junior Initiate to Phi Beta Kappa Honor Society
- Fall 2008 – Spring 2012: Dean’s List/Honors to Date in College of Letters and Sciences
Service and leadership
- 2024: Lead organizer, area chair and program chair of the ICLR 2024 Workshop on AI4DifferentialEquations In Science
- 2023-Present: Reviewer, International Conference on Learning Representations (ICLR)
- 2023-Present: Reviewer, SIAM Journal on Scientific Computing (SISC)
- 2023: Presented at Daves Avenue Elementary School Science Fair on “Mathematics in Science”
- 2022-Present: Reviewer, Nature Machine Intelligence, AISTATS
- 2021-Present: Reviewer, Conference on Neural Information Processing Systems (NeurIPS)
- 2019-Present: Reviewer, International Conference on Machine Learning (ICML)
- 2016-Present: Reviewer, Journal of Computational Physics, Science Advances
- 2013-Present: Technology, Engineering and Mathematical Sciences Judge
- Fall 2011: Panel Member for Letters & Science “Finding Your Way” Advising Program