Automated cognitive software application error detection

Patent metadata

Field	Value
Publication number	US-11429472-B1
Application number	US-202117301138-A
Country	US
Kind code	B1
Filing date	Mar 26, 2021
Priority date	Mar 26, 2021
Publication date	Aug 30, 2022
Grant date	Aug 30, 2022

Abstract

A method, system, and computer program product for implementing automated cognitive software application error detection is provided. The method includes receiving data associated with model based self-learning software code. The annotated data is automatically divided with respect to specified categorization and grouping attributes and categorized groups comprising portions of the annotated data are generated and analyzed. At least one incorrect annotation associated a group of the categorized groups is detected and filtered. Likewise, a correct annotation for the group is detected and retrieved from a database. The correct annotation is appended to the group.

First claim

What is claimed is:

1. An automated cognitive software application error detection method comprising:
receiving, by a processor of a server hardware device, annotated data associated with model based self-learning software code;
automatically dividing, by said processor, said annotated data with respect to specified categorization and grouping attributes;
generating, by said processor in response to results of said automatically dividing, categorized groups comprising portions of said annotated data;
analyzing, by said processor, said categorized groups;
detecting, by said processor based on results of said analyzing, at least one incorrect annotation associated with at least one group of said categorized groups;
filtering, by said processor, said at least one incorrect annotation from said at least one group of said categorized groups, wherein said filtering is executed based on a detected confidence level of said at least one incorrect annotation exceeding an error level threshold, and wherein generating said detected confidence level comprises:
selecting, by said processor, a target intent with respect to associated intents with respect to a vector matching process;
splicing, by said processor, a corpus within an intent associated with said target intent into an intent corpus;
transmitting, by said processor, said corpus within said intent corpus to a training model for obtaining a probability value specifying that said intent corpus is associated with said target intent, wherein said training model is executed for training a training set for generating predictions with respect to a code sample within a code test set of said model based self-learning software code; and
customizing, by said processor with respect to said predictions, said error level threshold based on a confidence level of said target intent;

detecting, by said processor within a database, a correct annotation for said at least one group of said categorized groups;
retrieving, by said processor from said database, said correct annotation;
appending, by said processor, said correct annotation to said at least one group of said categorized groups;
automatically executing, by said processor, error data rejection code with respect to code, of said model based self-learning software code, determined to be associated with an error;
highlighting, by said processor, said code with respect to marking errors resulting in highlighted code; and
reviewing, by said processor, said highlighted code for input into training code of said training model.

2. The method of claim 1 , wherein said annotated data is associated with artificial intelligence code of said model based self-learning software code.

3. The method of claim 1 , wherein said annotated data is associated with machine learning code of said model based self-learning software code.

4. The method of claim 1 , wherein each categorized group of said categorized groups comprises training set portions.

5. The method of claim 1 , wherein said automatically dividing said annotated data comprises:
adding Gaussian noises to a vector level for each data point of said annotated data within a same category;
generating peripheral vectors derived from each said data point; and
fitting newly generated support data to Gaussian mixture model software code.

6. The method of claim 1 , further comprising:
generating, by said processor, self-learning software code for executing future processes associated with executing said automated cognitive software application error detection method; and
storing, by said processor, said self-learning software code within a modified portion of a memory structure of said server hardware device.

7. The method of claim 6 , further comprising:
enabling, by said processor executing said self-learning software code, automated software and hardware control systems resulting in operation of hardware devices.

8. The method of claim 1 , further comprising:
providing at least one support service for at least one of creating, integrating, hosting, maintaining, and deploying computer-readable code in the server hardware device, said code being executed by the processor to implement: said receiving, said automatically dividing, said generating, said analyzing, said detecting said at least one incorrect annotation, said filtering, said detecting said correct annotation, said retrieving, and said appending.

9. A computer program product, comprising a computer readable hardware storage device storing a computer readable program code, said computer readable program code comprising an algorithm that when executed by a processor of a server hardware device implements an automated cognitive software application error detection method, said method comprising:
receiving, by said processor, annotated data associated with model based self-learning software code;
automatically dividing, by said processor, said annotated data with respect to specified categorization and grouping attributes;
generating, by said processor in response to results of said automatically dividing, categorized groups comprising portions of said annotated data;
analyzing, by said processor, said categorized groups;
detecting, by said processor based on results of said analyzing, at least one incorrect annotation associated with at least one group of said categorized groups;
filtering, by said processor, said at least one incorrect annotation from said at least one group of said categorized groups, wherein said filtering is executed based on a detected confidence level of said at least one incorrect annotation exceeding an error level threshold, and wherein generating said detected confidence level comprises:
selecting, by said processor, a target intent with respect to associated intents with respect to a vector matching process;
splicing, by said processor, a corpus within an intent associated with said target intent into an intent corpus;
transmitting, by said processor, said corpus within said intent corpus to a training model for obtaining a probability value specifying that said intent corpus is associated with said target intent, wherein said training model is executed for training a training set for generating predictions with respect to a code sample within a code test set of said model based self-learning software code; and
customizing, by said processor with respect to said predictions, said error level threshold based on a confidence level of said target intent;

detecting, by said processor within a database, a correct annotation for said at least one group of said categorized groups;
retrieving, by said processor from said database, said correct annotation;
appending, by said processor, said correct annotation to said at least one group of said categorized groups;
automatically executing, by said processor, error data rejection code with respect to code, of said model based self-learning software code, determined to be associated with an error;
highlighting, by said processor, said code with respect to marking errors resulting in highlighted code; and
reviewing, by said processor, said highlighted code for input into training code of said training model.

10. The computer program product of claim 9 , wherein said annotated data is associated with artificial intelligence code of said model based self-learning software code.

11. The computer program product of claim 9 , wherein said annotated data is associated with machine learning code of said model based self-learning software code.

12. The computer program product of claim 9 , wherein each categorized group of said categorized groups comprises training set portions.

Assignees

• IBM

Inventors

• Yuan Zhong Fang
• Liu Tong
• Zhang li ni
• Liang yong fang
• Gao Chen

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F2201/81

  Threshold · CPC title

• G06F11/3612 Primary

  by runtime analysis (performance monitoring G06F11/3466) · CPC title

• G06F11/0793

  Remedial or corrective actions (recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40) · CPC title