Unlock instant, AI-driven research and patent intelligence for your innovation.

Method for computer-implemented analysis of classification model

A classification model, computer technology, applied in computer parts, neural learning methods, computing, etc., can solve problems such as not working well

Pending Publication Date: 2020-10-27
SIEMENS AG
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Also, the method does not work well when many features of the classified instance support the classification

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
  • Method for computer-implemented analysis of classification model
  • Method for computer-implemented analysis of classification model
  • Method for computer-implemented analysis of classification model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] A deep neural network is a function that maps input features into a target space. Individual classification decisions of deep neural networks can be explained via saliency methods by creating saliency maps. However, the saliency maps produced by existing methods are not reliable. The two main model-agnostic interpretation methods are LIME and PDA.

[0030] LIME trains small classifiers (such as linear classifiers) with interpretable weights for approximating the local decision boundaries of deep classifiers. For each classification decision, this approach requires training a new classifier, which is not efficient. Given a single classification, the optimization of a classifier can end up with different parameters, which leads to inconsistent interpretations. When interpreting classifications by using different classifiers, LIME produces different interpretations.

[0031] PDA is another model-agnostic method for analyzing classifiers. PDA is a probabilistic acousti...

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 invention describes a method for computer-implemented analysis of a classification model which is adapted to map, as a prediction, a number of input instances, each of them having a number n of features, into a number of probabilities of output classes, as a classification decision, according to a predetermined function, and which is adapted to determine a relevance value for each feature resulting in a saliency map. The invention comprises the step of identifying (S1) an effect of each feature on the prediction of the instance by determining, for each feature, a relevance information representing a contextual information for all features of the instance omitting the considered feature. Then, the relevance value for each feature is determined (S2) by combining the relevance informationfor the features of the instance. Finally, the plurality of relevance values for the features of the instance is evaluated (S3) to identify the effect of each feature on the prediction of the instance.

Description

technical field [0001] The present invention relates to a method for computer-implemented analysis of a classification model adapted to map, as a prediction, a plurality of input instances, each of which has a plurality of features, according to a predetermined function into multiple probabilities of the output class as a classification decision. Furthermore, the classification model is adapted to determine a correlation value for each feature, which results in a saliency map. Background technique [0002] One of the weaknesses of machine learning models is the lack of interpretability. In particular, it is difficult in practice to explain intelligent decisions to end users. Models are often applied as black boxes, especially deep neural networks. Deep models achieve human-level performance for many tasks, such as image classification and detection. These neural networks with deep architectures are incomprehensible even to machine learning experts. However, intelligent ...

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): G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V10/462G06N3/045G06F18/241G06N5/045G06N7/01G06F16/55G06F16/5854G06F18/217G06F18/2431G06N3/047
Inventor 顾金东
Owner SIEMENS AG