John Dorelli - Heliophysics Applications of Physics Informed Machine Learning

preview_player
Показать описание
John Dorelli - Heliophysics Applications of Physics Informed Machine Learning

  1. Magnetosphere Seminars
Рекомендации по теме
John Dorelli - 0:43:45

John Dorelli - Heliophysics Applications of Physics Informed Machine Learning

Space MMS Science 0:03:35

Space MMS Science Overview The Many Mysteries of MMS

Zebang Shen, Entropy-dissipation 0:35:28

Zebang Shen, Entropy-dissipation Informed Neural Network for McKean-Vlasov Type PDEs

The Secrets Of 0:07:02

The Secrets Of Earth's Magnetic Field | The New Frontier Episode 8

NASA | MMS 0:03:36

NASA | MMS Science Overview: The Many Mysteries of MMS

ScienceCasts: NASA Spacecraft 0:04:56

ScienceCasts: NASA Spacecraft Fly in Record-setting Formation

2022 Heliophysics Summer 4:13:01

2022 Heliophysics Summer School August 10, 2022

Intro to Python 0:54:30

Intro to Python Notebooks and Code Sharing - Barbara Thompson and the Helioanalytics Group

Planetary Magnetosheres | 1:40:19

Planetary Magnetosheres | Fran Bagenal

Doc of the 0:23:48

Doc of the Day: Solar Weather Impact Report

The Magnetosphere as 1:05:10

The Magnetosphere as a System - Joe Borovsky

Stunning Facts About 0:22:35

Stunning Facts About the Sun You Need to Know. | Space: The New Frontier

Cole Dorman, Development 0:30:01

Cole Dorman, Development of Self-Calibrating Magnetoinductive Sensors for Spaceflight Constellations

Комментарии
Автор

Small question: When you talk about the model prediction having some small density errors within the Poission-errors (at about 30:35), is it possible that that is due to you having an error-function in the residuals derived from a Gaussian (at about 26:15)? Such a symmetric error-function might be sensitive to symmetries of the statistics of the measurements around the expected value - as would be the case for Poisson-distributed observations in the small-counts range...

KitagumaIgen
Автор

Hi, John. Greetings from Iowa. Would you happen to have a benchmark package that could be used to evaluate platforms for your MMS processing task? I know of a small business that's offering unlimited usage of limited cpu/gpu resources for $120/mo. with free samples.

Space-Audio