Learning based incident or defect resolution, and test generation

Patent metadata

Field	Value
Publication number	US-11233693-B2
Application number	US-202016926129-A
Country	US
Kind code	B2
Filing date	Jul 10, 2020
Priority date	Sep 15, 2017
Publication date	Jan 25, 2022
Grant date	Jan 25, 2022

Abstract

In some examples, learning based incident or defect resolution, and test generation may include ascertaining historical log data that includes incident or defect log data associated with operation of a process, and generating, based on the historical log data, step action graphs. Based on grouping of the step action graphs with respect to different incident and defect tickets, an incident and defect action graph may be generated to further generate a machine learning model. Based on an analysis of the machine learning model with respect to a new incident or defect, an output that includes a sequence of actions may be generated to reproduce, for the new incident, steps that result in the new incident, reproduce, for the new defect, an error that results in the new defect, identify a root cause of the new incident or defect, and/or resolve the new incident or defect.

First claim

What is claimed is:

1. A learning based incident or defect resolution, and test generation apparatus comprising:
a log data receiver, executed by at least one hardware processor, to ascertain historical log data that includes incident or defect log data associated with operation of a process;
a step action graph generator, executed by the at least one hardware processor, to generate, based on the historical log data, step action graphs;
an incident and defect action graph generator, executed by the at least one hardware processor, to generate, based on grouping of the step action graphs with respect to different incident and defect tickets, an incident and defect action graph for each of the incident and defect tickets; and
an output generator, executed by the at least one hardware processor, to generate, based on an analysis of a new incident or defect with respect to each of the incident and defect action graphs corresponding to the incident and defect tickets, an output to resolve the new incident or defect.

2. The learning based incident or defect resolution, and test generation apparatus according to claim 1 , wherein the step action graph generator is executed by the at least one hardware processor to generate, based on the historical log data, the step action graphs by:
analyzing, from the historical log data, data related to at least one of resolution of an incident or defect, or data related to reproduction of the incident or defect; and
transforming, from the analyzed data, textual details from each step into the step action graphs by identifying
action terms,
subjects initiating the action terms, and
objects associated with the actions terms.

3. The learning based incident or defect resolution, and test generation apparatus according to claim 1 , wherein the step action graph generator is executed by the at least one hardware processor to generate, based on the historical log data, the step action graphs by:
identifying adjacent action terms; and
joining nodes related to the adjacent action terms by sequential flow edges.

4. The learning based incident or defect resolution, and test generation apparatus according to claim 1 , wherein the step action graph generator is executed by the at least one hardware processor to generate, based on the historical log data, the step action graphs by:
identifying conditionally related action terms; and
joining nodes related to the conditionally related action terms by conditional branching.

5. The learning based incident or defect resolution, and test generation apparatus according to claim 1 , wherein the incident and defect action graph generator is executed by the at least one hardware processor to generate, based on grouping of the step action graphs with respect to different incident and defect tickets, the incident and defect action graph for each of the incident and defect tickets by:
analyzing, for the step action graphs, directed flow edges that identify sequences of actions; and
generating, based on grouping of the step action graphs according to the directed flow edges that identify the sequences of actions and with respect to the different incident and defect tickets, the incident and defect action graph for each of the incident and defect tickets.

6. The learning based incident or defect resolution, and test generation apparatus according to claim 1 , further comprising:
a machine learning model generator, executed by the at least one hardware processor, to generate a machine learning model by:
learning non-linear patterns of actions to resolve an incident or defect from the historical log data.

7. The learning based incident or defect resolution, and test generation apparatus according to claim 1 , wherein the output generator is executed by the at least one hardware processor to generate, based on the analysis of the new incident or defect with respect to each of the incident and defect action graphs corresponding to the incident and defect tickets, the output to resolve the new incident or defect by:
determining semantic relatedness between the new incident or defect, and incidents or defects, respectively, of the historical log data; and
determining whether selection of a step from the historical log data causally constrains selection of a subsequent step from the historical log data.

8. The learning based incident or defect resolution, and test generation apparatus according to claim 7 , wherein the output generator is executed by the at least one hardware processor to determine whether selection of the step from the historical log data causally constrains selection of the subsequent step from the historical log data by:
determining a number of sequences where the step from the historical log data precedes the subsequent step from the historical log data; and
assigning, based on the number of sequences, a degree of confidence to the step from the historical log data that precedes the subsequent step from the historical log data.

9. The learning based incident or defect resolution, and test generation apparatus according to claim 1 , wherein the output generator is executed by the at least one hardware processor to generate, based on the analysis of the new incident or defect with respect to each of the incident and defect action graphs corresponding to the incident and defect tickets, the output to resolve the new incident or defect by:
analyzing step to defect similarity between a step of the historical log data and the new incident or defect; and
analyzing defect to defect similarity between a defect of the historical log data and the new incident or defect.

10. The learning based incident or defect resolution, and test generation apparatus according to claim 1 , further comprising:
a user feedback analyzer, executed by the at least one hardware processor, to
identify, in the generated output, an incomplete feature of the generated output;
ascertain feedback with respect to the incomplete feature;
update, based on the feedback, the historical log data with respect to the incomplete feature; and
update a machine learning model based on an analysis of the updated historical log data.

11. The learning based incident or defect resolution, and test generation apparatus according to claim 10 , wherein the user feedback analyzer is executed by the at least one hardware processor to:
generate a further incident and defect action graph with respect to the new incident or defect;
identify, with respect to the incomplete feature, a difference in an existing incident and defect action graph and the further incident and defect action graph; and
continue iterations related to the update of the historical log data with respect to a further incomplete feature and the update of the machine learning model with respect to a further update to the historical log data until elimination of differences in existing incident and defect action graphs and further incident and defect action graphs related to the new incident or defect.

12. A computer implemented method for learning based incident or defect resolution, and test generation comprising:
ascertaining, by at least one hardware processor, historical log data that includes existing tests;
generating, by the at least one hardware processor, based on the historical log data, step action graphs;
generating, by the at least one hardware processor, based on grouping of the step action graphs with respect to different tests, a test action graph for each of the different tests; and
identifying, by the at least one hardware processor, based on an analysis of the historical log data with respect to each of the test action graphs correspondi

Assignees

• Accenture Global Solutions Ltd

Inventors

• Misra Janardan
• Rawat Divya
• Sengupta Shubhashis

Classifications

• G06N7/01

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

• H04L41/0645

  by additionally acting on or stimulating the network after receiving notifications · CPC title

• G06F11/3476

  Data logging (G06F11/14, G06F11/2205 take precedence) · CPC title

• H04L41/16

  using machine learning or artificial intelligence · CPC title

• G06F11/3684

  for test design, e.g. generating new test cases · CPC title