System and method for assisting user to resolve a hardware issue and a software issue

Patent metadata

Field	Value
Publication number	US-10769043-B2
Application number	US-201816017402-A
Country	US
Kind code	B2
Filing date	Jun 25, 2018
Priority date	Jun 25, 2018
Publication date	Sep 8, 2020
Grant date	Sep 8, 2020

Abstract

The present disclosure relates to system(s) and method(s) for assisting a user to resolve a hardware issue and a software issue. The system identifies, a target cluster, associated with a new ticket received from the user, from the set of clusters. Further, the system recommends one or more runbook scripts, from a runbook repository, associated with the new ticket. The system further identifies a new runbook script, corresponding to the new ticket, from a set of external repositories. Further, the system executes at least one of the one or more runbook scripts or the new runbook script, associated with the new ticket. The system further generates a document based on the execution of the one or more runbook scripts or the new runbook script, thereby assisting the user to resolve a target issue.

First claim

We claim:

1. A system for assisting a user to resolve a hardware issue and a software issue, the system comprising:
a memory;
a processor coupled to the memory, wherein the processor is configured to execute programmed instructions stored in the memory to:
categorise a set of tickets stored in a ticket repository to generate a set of clusters, wherein each cluster, from the set of clusters, is configured to maintain one or more tickets from the set of tickets, wherein each ticket is associated with a ticket description, and wherein the categorization is based on the ticket description;
identify a target cluster, from the set of clusters, associated with a new ticket received from a user, wherein a new ticket description, associated with the new ticket, is analysed using a Convolution Neural Network based neural language model for identifying the target cluster, and wherein the new ticket corresponds to a target issue;
recommend one or more runbook scripts, from a runbook repository, associated with the new ticket based on analysis of the new ticket description using an artificial intelligence based recommendation engine, wherein the one or more runbook scripts are associated with a runbook script description, and wherein the one or more runbook scripts comprise a probability, to resolve the new ticket, higher than a pre-defined threshold probability;
identify a new runbook script, corresponding to the new ticket, from a set of external repositories using a natural language processing and a text extraction algorithm, when the one or more runbook scripts are not available in the runbook repository or the one or more runbook scripts fail to resolve the target issue;
execute at least one of the one or more runbook scripts or the new runbook script associated with the new ticket; and
generate a document based on the execution of the one or more runbook scripts or the new runbook script, thereby assisting the user to resolve the target issue.

2. The system of claim 1 , further configured to generate the new runbook script, corresponding to the new ticket, using a natural language processing and a reinforcement learning, wherein the new runbook script is generated based on analysis of information extracted from the set of external repositories.

3. The system of claim 1 , wherein each ticket from the set of tickets is associated with one of a hardware issue or a software issue.

4. The system of claim 1 , wherein the document comprises the new ticket, the new ticket description, the target cluster, the one or more runbook scripts, and the new runbook script.

5. The system of claim 1 , further configured to store the new runbook script, associated with the new ticket, in the runbook repository.

6. The system of claim 1 , further configured to process the ticket description associated with each ticket based on:
parsing the ticket description, corresponding to each ticket from the set of tickets, to generate a set of tokens;
removing one or more stop words from the ticket description of each ticket based on analysis of the set of tokens;
extracting one or more words from the ticket description of each ticket based on the removal of one or more stop words; and
lemmatizing the one or more words of the ticket description associated with each ticket, wherein each ticket from the set of ticket is categorised based on the lemmatization.

7. The system of claim 1 , further configured to compute a Q-value for a new ticket, wherein the Q-value is used to identify the target cluster associated with the new ticket.

8. The system of claim 1 , further configured to map a set of runbook scripts, stored in the runbook repository, to the set of tickets using a Convolution Neural Network based neural language model.

9. The system of claim 1 , wherein the one or more runbook scripts comprises a runbook name, a runbook tag, a runbook description, and category.

10. The system of claim 1 , further configured to prioritise the one or more runbook scripts based on:
extracting the runbook script description associated with the one or more runbook scripts from the runbook repository;
converting the runbook script description into a condensed vector using deep auto encoders;
computing a similarity between the new ticket description and the runbook script description;
prioritising the one or more runbook scripts based on the similarity between the new ticket description and the runbook script description, wherein the similarity between the new ticket description and the runbook script description is computed using a neural language model.

11. The system of claim 1 , wherein the target cluster is identified based on the analysis of the new ticket description based on:
generating a vector for the new ticket description associated with the new ticket based on concatenating one or more vectors associated with each word from the new ticket description;
extracting one or more features from the new ticket description based on identification of one or more phrases on the new ticket description;
generating a feature vector based on execution of a max pooling operation on the one or more features; and
connecting the feature vectors associated with the one or more features, wherein the feature vectors are used to identify the target cluster.

12. A method for assisting a user to resolve a hardware issue and a software issue, the method comprising steps of:
categorising, by a processor, a set of tickets stored in a ticket repository to generate a set of clusters, wherein each cluster, from the set of clusters, is configured to maintain one or more tickets from the set of tickets, wherein each ticket is associated with a ticket description, and wherein the categorization is based on the ticket description;
identifying, by the processor, a target cluster, from the set of clusters, associated with a new ticket received from a user, wherein a new ticket description, associated with the new ticket, is analysed using a Convolution Neural Network based neural language model for identifying the target cluster, and wherein the new ticket corresponds to a target issue;
recommending, by the processor, one or more runbook scripts, from a runbook repository, associated with the new ticket based on analysis of the new ticket description using an artificial intelligence based recommendation engine, wherein the one or more runbook scripts are associated with a runbook script description, and wherein the one or more runbook scripts comprise a probability, to resolve the new ticket, higher than a pre-defined threshold probability;
identifying, by the processor, a new runbook script, corresponding to the new ticket, from a set of external repositories using a natural language processing and a text extraction algorithm, when the one or more runbook scripts are not available in the runbook repository or the one or more runbook scripts fail to resolve the target issue;
executing, by the processor, at least one of the one or more runbook scripts or the new runbook script associated with the new ticket; and
generating, by the processor, a document based on the execution of the one or more runbook scripts or the new runbook script, thereby assisting the user to resolve the target issue.

13. The method of claim 12 , further comprises generation of the new runbook script, corresponding to the new ticket, using a natural language processing and a reinforcement learning, wherein the new runbook script is generated based on analysis of information extracted from the set of external repositories.

14. The method of claim 12 , wherein each ticket from the set of tickets is associated with one of

Assignees

• Hcl Technologies Ltd
• Hcl Technologies Ltd

Inventors

• Sabharwal Navin
• Agrawal Amit

Classifications

• G06N20/00 Primary

  Machine learning · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title

• G06N3/092

  Reinforcement learning · CPC title