Oak Ridge National Laboratories / Artificial Neural Networks / Ford / S & P / SPSS / /
Currency
USD / / /
Facility
College Park / University of Maryland / /
IndustryTerm
energy validation report / action network / information processing / integrated software system / larger models and modeling systems / ordered computational systems / neural networks / feedforward network / greater computing capacity / social communications / implicit systems / recurrent network / data mining / natural gas industry / software package / real-world applications / purpose software package / universal neural network learning device / natural gas regulation / computer systems / neural network / multivariate statistical tool / recurrent networks / actual physical infrastructure / elementary processing elements / explicit computational algorithm / natural gas markets / control systems / recursive neural networks / feedforward networks / predicted residential gas price / technology applications / Web From Sensors / distributed web / ordered systems / larger systems / /
Organization
National Science Foundation / MIT / Harvard / Department of Energy / University of Maryland / College Park / Energy Information Administration / office of Energy Information Validation / IEEE Neural Network Society / Engineering Directorate / US Federal Reserve / Department of Homeland Security / Department of Defense / Defense Advanced Projects Agency / /
Person
Louis Rall / Robert Newcomb / Modelof / Marvin Minsky / Karl Deutsch / / /
Position
author / not the official views / mathematician / Governor / Model / Professor / intelligent controller / official views / /
Backwards Differentiation in AD and Neural Nets: Past Links and New Opportunities Paul J. Werbos1 Abstract Backwards calculation of derivatives – sometimes called the reverse mode, the full adjoint method, or backpropa