Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Relevance score assignment for artificial neural network

An artificial neural network, correlation score technology, applied in biological neural network models, neural architectures, neural learning methods, etc., can solve problems such as preventing human experts from carefully verifying classification decisions

Active Publication Date: 2018-01-26
FRAUNHOFER GESELLSCHAFT ZUR FOERDERUNG DER ANGEWANDTEN FORSCHUNG EV +1
View PDF2 Cites 21 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This is a considerable drawback in classification applications as it prevents human experts from carefully validating classification decisions

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Relevance score assignment for artificial neural network
  • Relevance score assignment for artificial neural network
  • Relevance score assignment for artificial neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] Before describing various embodiments of the present application with respect to block diagrams, the concepts underlying the concepts of the embodiments are first described by briefly introducing artificial neural networks and then explaining the concepts underlying the concepts of the embodiments.

[0033]A neural network is a graph of interconnected nonlinear processing units that can be trained to approximate complex mappings between input and output data. Note that the input data is e.g. an image (set of pixels) and the output is e.g. a classification decision (in the simplest case, +1 / -1 means "yes" there is a shark in the image or "no" there is no shark in the image). Each nonlinear processing unit (or neuron) consists of a weighted linear combination of its inputs to which a nonlinear activation function is applied. Using index i to denote each neuron entering the neuron with index j, the nonlinear activation function is defined as:

[0034]

[0035] where g(...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The task of relevance score assignment to a set of items onto which an artificial neural network is applied is obtained by redistributing an initial relevance score derived from the network output, onto the set of items by reversely propagating the initial relevance score through the artificial neural network so as to obtain a relevance score for each item. In particular, this reverse propagationis applicable to a broader set of artificial neural networks and / or at lower computational efforts by performing same in a manner so that for each neuron, preliminarily redistributed relevance scoresof a set of downstream neighbor neurons of the respective neuron are distributed on a set of upstream neighbor neurons of the respective neuron according to a distribution function.

Description

technical field [0001] This application is concerned with relevance score assignment for artificial neural networks. Such relevance score assignments can be used, for example, for region of interest (ROI) identification. Background technique [0002] Computer programs are able to successfully solve many complex tasks, such as the automatic classification of images and text or the assessment of a person's reputation. Machine learning algorithms are especially successful because they learn from data, i.e. the program gets a large labeled (or weakly labeled) training set, and after a certain training phase, it is able to generalize to new unseen examples. Many banks have systems for categorizing the creditworthiness of people applying for loans (eg, based on age, address, income, etc.). The main disadvantage of such systems is explainability, i.e. the system usually does not provide information on why and how a decision was made (e.g. why someone is classified as not reputabl...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06K9/46G06N3/04G06N3/08
CPCG06N3/04G06N3/084G06V10/454G06V10/7715G06N3/08G06T2207/20084G06F18/2135G06F40/279G10L25/30G06N3/02G06N3/048
Inventor 塞巴斯蒂安·巴赫沃耶西·萨梅克克劳斯-罗伯特·穆勒亚历山大·宾德格雷格里·蒙塔翁
Owner FRAUNHOFER GESELLSCHAFT ZUR FOERDERUNG DER ANGEWANDTEN FORSCHUNG EV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com