Automated compilation of probabilistic task description into executable neural network specification

Simple item page
dc.contributor.assigneeRice Universityen_US
dc.contributor.publisherUnited States Patent and Trademark Officeen_US
dc.creatorPatel, Ankit B.en_US
dc.creatorBaraniuk, Richard G.en_US
dc.date.accessioned2021-03-02T23:16:38Zen_US
dc.date.available2021-03-02T23:16:38Zen_US
dc.date.filed2016-03-11en_US
dc.date.issued2020-11-24en_US
dc.description.abstractA mechanism for compiling a generative description of an inference task into a neural network. First, an arbitrary generative probabilistic model from the exponential family is specified (or received). The model characterizes a conditional probability distribution for measurement data given a set of latent variables. A factor graph is generated for the generative probabilistic model. Each factor node of the factor graph is expanded into a corresponding sequence of arithmetic operations, based on a specified inference task and a kind of message passing algorithm. The factor graph and the sequences of arithmetic operations specify the structure of a neural network for performance of the inference task. A learning algorithm is executed, to determine values of parameters of the neural network. The neural network is then ready for performing inference on operational measurements.en_US
dc.digitization.specificationsThis patent information was downloaded from the US Patent and Trademark website (http://www.uspto.gov/) as image-PDFs. The PDFs were OCRed for access purposesen_US
dc.format.extent54en_US
dc.identifier.citationPatel, Ankit B. and Baraniuk, Richard G., "Automated compilation of probabilistic task description into executable neural network specification." Patent US10846589B2. issued 2020-11-24. Retrieved from <a href="https://hdl.handle.net/1911/110123">https://hdl.handle.net/1911/110123</a>.en_US
dc.identifier.patentIDUS10846589B2en_US
dc.identifier.urihttps://hdl.handle.net/1911/110123en_US
dc.language.isoengen_US
dc.titleAutomated compilation of probabilistic task description into executable neural network specificationen_US
dc.typeUtility patenten_US
dc.type.dcmiTexten_US
dc.type.genrepatentsen_US
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
US10846589B2.pdf
Size:
3.18 MB
Format:
Adobe Portable Document Format
Description:
Download
Collections
Rice Patents