IMA UMN
IMA UMN
  • Видео 158
  • Просмотров 86 388
On New Results on Ensemble Methods for Learning, Optimization, and Filtering – Michael Herty
IMA Data Science Seminar
Speaker: Michael Herty (RWTH Aachen University)
"On New Results on Ensemble Methods for Learning, Optimization, and Filtering"
Abstract: We are interested in the construction of numerical methods for (constrained) very high-dimensional constrained nonlinear optimization problems by gradient free techniques appearing data assimilation problems as well as learning applications. Gradients are replaced by particle approximations and recently different methods have been proposed, e.g. consensus-based, swarm-based or ensemble Kalman based methods in case of filtering. We discuss new results on constrained and parametric case as well as their corresponding mean field descri...
Просмотров: 30

Видео

Flag manifolds for robust averaging, principal directions, and hierarchical data representation55:31Flag manifolds for robust averaging, principal directions, and hierarchical data representation
Flag manifolds for robust averaging, principal directions, and hierarchical data representation
Просмотров 32День назад
IMA Data Science Seminar Speaker: Nathan Mankovich (University of Valencia) "Flag manifolds for robust averaging, principal directions, and hierarchical data representation" Abstract:A flag is a sequence of nested subspaces of a finite-dimensional vector space and the flag manifold is the collection of all flags with the same sequence of subspace dimensions. Flag manifolds are used in computer ...
Graph-based Semi-supervised and Unsupervised Local Clustering - Zhaiming Shen55:46Graph-based Semi-supervised and Unsupervised Local Clustering - Zhaiming Shen
Graph-based Semi-supervised and Unsupervised Local Clustering - Zhaiming Shen
Просмотров 43День назад
IMA Data Science Seminar Speaker: Zhaiming Shen (Georgia Tech) "Graph-based Semi-supervised and Unsupervised Local Clustering" Abstract: Semi-supervised local clustering focuses on identifying specific substructures within a large graph without requiring a full understanding of the entire graph. In this talk, I will present a method to find the target cluster by solving a sparse solution to a l...
On efficient, approximate sampling for high dimensional scientific computing - Yifan Chen1:03:45On efficient, approximate sampling for high dimensional scientific computing - Yifan Chen
On efficient, approximate sampling for high dimensional scientific computing - Yifan Chen
Просмотров 83День назад
IMA Data Science Seminar Speaker: Yifan Chen "On efficient, approximate sampling for high dimensional scientific computing" Abstract:Estimating and sampling from high-dimensional probability distributions is a fundamental yet challenging task in scientific computing and machine learning. Approximate, biased algorithms appear empirically more efficient and scalable than unbiased, exact methods. ...
From academia to Industry. Old and new challenges and strategies. - Luis Duque Alvarez1:00:13From academia to Industry. Old and new challenges and strategies. - Luis Duque Alvarez
From academia to Industry. Old and new challenges and strategies. - Luis Duque Alvarez
Просмотров 18День назад
IMA Industrial Problems Seminar Speaker: Luis Duque Alvarez (Meta) "From academia to Industry. Old and new challenges and strategies." Abstract: In this talk I will talk about the preparation, networking and overall strategies that I used when transitioning from academia to industry and will also talk about experiences of colleagues that did this transition recently. Presented virtually for the...
In-context learning of solution operators to linear elliptic PDEs - Frank Cole59:43In-context learning of solution operators to linear elliptic PDEs - Frank Cole
In-context learning of solution operators to linear elliptic PDEs - Frank Cole
Просмотров 72День назад
IMA Data Science Seminar Speaker: Frank Cole (University of Minnesota) "In-context learning of solution operators to linear elliptic PDEs" Abstract:Transformer-based neural networks have demonstrated the ability learn in-context: given a few demonstrations of a new task, they can make correct predictions without updating their parameters. Recently, transformer-based foundation models have emerg...
Universally counting curves in Calabi-Yau threefolds - John Pardon1:01:02Universally counting curves in Calabi-Yau threefolds - John Pardon
Universally counting curves in Calabi-Yau threefolds - John Pardon
Просмотров 103День назад
Universally counting curves in Calabi-Yau threefolds - John Pardon (Simons Center for Geometry and Physics) 2024 Yamabe Symposium at the University of Minnesota: cse.umn.edu/math/yamabe-memorial-symposium
Symplectic excision - Yael Karshon1:09:18Symplectic excision - Yael Karshon
Symplectic excision - Yael Karshon
Просмотров 47День назад
Symplectic excision - Yael Karshon (Tel-Aviv University and University of Toronto) 2024 Yamabe Symposium at the University of Minnesota: cse.umn.edu/math/yamabe-memorial-symposium
Two or infinity - Daniel Cristofaro-Gardiner1:07:27Two or infinity - Daniel Cristofaro-Gardiner
Two or infinity - Daniel Cristofaro-Gardiner
Просмотров 23День назад
Two or infinity - Daniel Cristofaro-Gardiner (University of Maryland, College Park) 2024 Yamabe Symposium at the University of Minnesota: cse.umn.edu/math/yamabe-memorial-symposium
Beyond reflexivity, symplectically - Margaret Symington57:18Beyond reflexivity, symplectically - Margaret Symington
Beyond reflexivity, symplectically - Margaret Symington
Просмотров 12День назад
Beyond reflexivity, symplectically - Margaret Symington (Mercer University) 2024 Yamabe Symposium at the University of Minnesota: cse.umn.edu/math/yamabe-memorial-symposium
Symplectic rational homology ball fillings of Seifert fibered spaces - John Etnyre54:09Symplectic rational homology ball fillings of Seifert fibered spaces - John Etnyre
Symplectic rational homology ball fillings of Seifert fibered spaces - John Etnyre
Просмотров 4314 дней назад
Symplectic rational homology ball fillings of Seifert fibered spaces - John Etnyre (Georgia Institute of Technology) 2024 Yamabe Symposium at the University of Minnesota: cse.umn.edu/math/yamabe-memorial-symposium
Floer homotopy as a bordism theory - Mohammed Abouzaid1:06:51Floer homotopy as a bordism theory - Mohammed Abouzaid
Floer homotopy as a bordism theory - Mohammed Abouzaid
Просмотров 10714 дней назад
Floer homotopy as a bordism theory - Mohammed Abouzaid (Stanford University) 2024 Yamabe Symposium at the University of Minnesota: cse.umn.edu/math/yamabe-memorial-symposium
Disk counting for tropical Lagrangians - Chris Woodward58:29Disk counting for tropical Lagrangians - Chris Woodward
Disk counting for tropical Lagrangians - Chris Woodward
Просмотров 2014 дней назад
Disk counting for tropical Lagrangians - Chris Woodward (Rutgers University, New Brunswick) 2024 Yamabe Symposium at the University of Minnesota: cse.umn.edu/math/yamabe-memorial-symposium
A Morse A∞-model for the higher-dimensional Heegaard Floer homology of cotangent fibers - Ko Honda1:07:18A Morse A∞-model for the higher-dimensional Heegaard Floer homology of cotangent fibers - Ko Honda
A Morse A∞-model for the higher-dimensional Heegaard Floer homology of cotangent fibers - Ko Honda
Просмотров 4814 дней назад
A Morse A∞-model for the higher-dimensional Heegaard Floer homology of cotangent fibers - Ko Honda (University of California, Los Angeles) 2024 Yamabe Symposium at the University of Minnesota: cse.umn.edu/math/yamabe-memorial-symposium
Generative Machine Learning Models for Uncertainty Quantification - Guannan Zhang1:08:02Generative Machine Learning Models for Uncertainty Quantification - Guannan Zhang
Generative Machine Learning Models for Uncertainty Quantification - Guannan Zhang
Просмотров 5684 месяца назад
IMA Data Science Seminar Speaker: Guannan Zhang (Oak Ridge National Laboratory) "Generative Machine Learning Models for Uncertainty Quantification" Abstract: Generative machine learning models, including variational auto-encoders (VAE), normalizing flows (NF), generative adversarial networks (GANs), diffusion models, have dramatically improved the quality and realism of generated content, wheth...
Effect of Leaky ReLUs on the training & generalization of overparameterized networks - Yinglong Guo53:56Effect of Leaky ReLUs on the training & generalization of overparameterized networks - Yinglong Guo
Effect of Leaky ReLUs on the training & generalization of overparameterized networks - Yinglong Guo
Просмотров 2204 месяца назад
Effect of Leaky ReLUs on the training & generalization of overparameterized networks - Yinglong Guo
Local geometry determines global landscape in low-rank - Shuyang Ling54:45Local geometry determines global landscape in low-rank - Shuyang Ling
Local geometry determines global landscape in low-rank - Shuyang Ling
Просмотров 1606 месяцев назад
Local geometry determines global landscape in low-rank - Shuyang Ling
Are the measurement data enough - Hongkai Zhao1:03:34Are the measurement data enough - Hongkai Zhao
Are the measurement data enough - Hongkai Zhao
Просмотров 646 месяцев назад
Are the measurement data enough - Hongkai Zhao
Spatial Stochastic modeling for population dynamic - Wai-Ton (Louis) Fan55:43Spatial Stochastic modeling for population dynamic - Wai-Ton (Louis) Fan
Spatial Stochastic modeling for population dynamic - Wai-Ton (Louis) Fan
Просмотров 846 месяцев назад
Spatial Stochastic modeling for population dynamic - Wai-Ton (Louis) Fan
Advancing Machine-Learned Interatomic Potentials: Enhancing Accuracy & Robustness in Materials Sci.52:07Advancing Machine-Learned Interatomic Potentials: Enhancing Accuracy & Robustness in Materials Sci.
Advancing Machine-Learned Interatomic Potentials: Enhancing Accuracy & Robustness in Materials Sci.
Просмотров 1476 месяцев назад
Advancing Machine-Learned Interatomic Potentials: Enhancing Accuracy & Robustness in Materials Sci.
On small & large scales in training physics-informed neural networks for PDEs - Zhongqiang Zhang1:03:07On small & large scales in training physics-informed neural networks for PDEs - Zhongqiang Zhang
On small & large scales in training physics-informed neural networks for PDEs - Zhongqiang Zhang
Просмотров 1336 месяцев назад
On small & large scales in training physics-informed neural networks for PDEs - Zhongqiang Zhang
Viva la Revolución of Open Source Large Language Models: Unleashing the Dark Horse in AI Innovation58:05Viva la Revolución of Open Source Large Language Models: Unleashing the Dark Horse in AI Innovation
Viva la Revolución of Open Source Large Language Models: Unleashing the Dark Horse in AI Innovation
Просмотров 417 месяцев назад
Viva la Revolución of Open Source Large Language Models: Unleashing the Dark Horse in AI Innovation
Random High-Dimensional Binary Vectors, Kernel Methods, & Hyperdimensional Computing - N. Marshall49:54Random High-Dimensional Binary Vectors, Kernel Methods, & Hyperdimensional Computing - N. Marshall
Random High-Dimensional Binary Vectors, Kernel Methods, & Hyperdimensional Computing - N. Marshall
Просмотров 4277 месяцев назад
Random High-Dimensional Binary Vectors, Kernel Methods, & Hyperdimensional Computing - N. Marshall
Improved Convergence Rates of Anderson Acceleration for a Large Class of Fixed-Point Iterations1:09:42Improved Convergence Rates of Anderson Acceleration for a Large Class of Fixed-Point Iterations
Improved Convergence Rates of Anderson Acceleration for a Large Class of Fixed-Point Iterations
Просмотров 3938 месяцев назад
Improved Convergence Rates of Anderson Acceleration for a Large Class of Fixed-Point Iterations
Generalization theory for diffusion models - Frank Cole56:20Generalization theory for diffusion models - Frank Cole
Generalization theory for diffusion models - Frank Cole
Просмотров 3918 месяцев назад
Generalization theory for diffusion models - Frank Cole
Optimal Transport Maps for Conditional Simulation - Bamdad Hosseini1:04:00Optimal Transport Maps for Conditional Simulation - Bamdad Hosseini
Optimal Transport Maps for Conditional Simulation - Bamdad Hosseini
Просмотров 2238 месяцев назад
Optimal Transport Maps for Conditional Simulation - Bamdad Hosseini
How much math do you really need to make markets in stock options? - John Dodson1:25:49How much math do you really need to make markets in stock options? - John Dodson
How much math do you really need to make markets in stock options? - John Dodson
Просмотров 1479 месяцев назад
How much math do you really need to make markets in stock options? - John Dodson
Learning in the presence of low-dimensional structure: a spiked random matrix perspective - Denny Wu58:42Learning in the presence of low-dimensional structure: a spiked random matrix perspective - Denny Wu
Learning in the presence of low-dimensional structure: a spiked random matrix perspective - Denny Wu
Просмотров 1,1 тыс.11 месяцев назад
Learning in the presence of low-dimensional structure: a spiked random matrix perspective - Denny Wu
The Kagome lattice as a mechanism-based mechanical metamaterial - Xuenan Li1:00:45The Kagome lattice as a mechanism-based mechanical metamaterial - Xuenan Li
The Kagome lattice as a mechanism-based mechanical metamaterial - Xuenan Li
Просмотров 31911 месяцев назад
The Kagome lattice as a mechanism-based mechanical metamaterial - Xuenan Li
Normalization effects and mean field theory for deep neural networks - Konstantinos Spiliopoulos1:06:36Normalization effects and mean field theory for deep neural networks - Konstantinos Spiliopoulos
Normalization effects and mean field theory for deep neural networks - Konstantinos Spiliopoulos
Просмотров 37411 месяцев назад
Normalization effects and mean field theory for deep neural networks - Konstantinos Spiliopoulos