Centralized machine learning predictor for a remote network management platform

Patent metadata

Field	Value
Publication number	US-11595484-B2
Application number	US-201916402800-A
Country	US
Kind code	B2
Filing date	May 3, 2019
Priority date	May 3, 2019
Publication date	Feb 28, 2023
Grant date	Feb 28, 2023

Abstract

A remote network management platform is provided that includes an end-user computational instance dedicated to a managed network, a training computational instance, and a prediction computational instance. The training instance is configured to receive a corpus of textual records from the end-user instance and to determine therefrom a machine learning (ML) model to determine the numerical similarity between input textual records and textual records in the corpus of textual records. The prediction instance is configured to receive the ML model and an additional textual record from the end-user instance, to use the ML model to determine respective numerical similarities between the additional textual record and the textual records in the corpus of textual records, and to transmit, based on the respective numerical similarities, representations of one or more of the textual records in the corpus of textual records to the end-user computational instance.

First claim

What is claimed is:

1. A remote network management platform comprising:
an end-user computational instance comprising a first set of computational resources of the remote network management platform and dedicated to a managed network;
a training computational instance comprising a second set of computational resources of the remote network management platform, and configured to perform operations including:
receiving, from the end-user computational instance, a corpus of textual records;
training, based on the corpus of textual records, a machine learning (ML) model to:
determine, for a cluster of textual records in the corpus of textual records, an aggregate vector representation in a vector space;
determine, for an input textual record, a vector representation in the vector space; and
determine a degree of numerical similarity between the input textual record and the cluster of textual records in the corpus of textual records by determining an angle, a distance, or a combination thereof, within the vector space, between the vector representation of the input textual record and the aggregate vector representation; and

transmitting, to the end-user computational instance, the ML model; and

a prediction computational instance comprising a third set of computational resources of the remote network management platform, and configured to perform operations including:
receiving, from the end-user computational instance, an additional textual record;
receiving, from the end-user computational instance, the ML model;
determining, by the ML model, respective numerical similarities between the additional textual record and the cluster of textual records in the corpus of textual records; and
based on the respective numerical similarities, transmitting, to the end-user computational instance, representations of one or more textual records in the corpus of textual records.

2. The remote network management platform of claim 1 , wherein the third set of computational resources includes more memory than the first set of computational resources.

3. The remote network management platform of claim 1 , wherein the third set of computational resources includes a memory, and wherein the operations performed by the prediction computational instance include:
in response to receiving the additional textual record, determining that the ML model is not stored in the memory; and
requesting, from the end-user computational instance, the ML model.

4. The remote network management platform of claim 1 , wherein determining, by the ML model, respective numerical similarities between the additional textual record and the cluster of the textual records in the corpus of textual records comprises using the ML model to determine at least one of (i) word vectors that describe, in a first semantically-encoded vector space, a meaning of respective words of the additional textual record, or (ii) a paragraph vector that describes, in a second semantically-encoded vector space, a meaning of multiple words of the additional textual record.

5. The remote network management platform of claim 4 , wherein the ML model represents a set of clusters of the textual records in the corpus of textual records, and wherein determining, by the ML model, respective numerical similarities between the additional textual record and the cluster of textual records in the corpus of textual records comprises selecting, for the additional textual record, the cluster from the set of clusters based on at least one of the word vectors or the paragraph vector.

6. The remote network management platform of claim 1 , wherein the ML model represents a set of clusters of textual records in the corpus of textual records, and wherein determining, by the ML model, respective numerical similarities between the additional textual record and the cluster of textual records in the corpus of textual records comprises selecting, for the additional textual record, the cluster from the set of clusters.

7. The remote network management platform of claim 1 , wherein the operations performed by the training computational instance include:
training the ML model to:
determine, for a given textual record in the corpus of textual records, a respective vector representation in the vector space; and
determine the degree of numerical similarity between the input textual record and the given textual record by determining a distance, within the vector space, between the vector representations of the input textual record and the given textual record.

8. The remote network management platform of claim 7 , wherein the operations performed by the training computational instance include training the ML model to
determine the degree of numerical similarity between the input textual record and the given textual record by determining an additional angle, within the vector space, between the vector representations of the input textual record and the given textual record.

9. The remote network management platform of claim 1 , wherein the operations performed by the training computational instance include training the ML model to:
determine, for the cluster of textual records in the corpus of textual records, a representative volume within the vector space;
and
determine the degree of numerical similarity between the input textual record and the cluster of textual records in the corpus of textual records by determining whether the vector representation of the input textual record is within the representative volume.

10. The remote network management platform of claim 1 , comprising:
an additional end-user computational instance comprising a fourth set of computational resources of the remote network management platform and dedicated to an additional managed network;
wherein the operations performed by the training computational instance include:
receiving, from the additional end-user computational instance, an additional corpus of textual records,
training, based on the additional corpus of textual records, an additional ML model to determine an additional degree of similarity between input textual records and textual records in the additional corpus of textual records, and
transmitting, to the additional end-user computational instance, the additional ML model; and

wherein the operations performed by the prediction computational instance include:
receiving, from the additional end-user computational instance, a further textual record,
receiving, from the additional end-user computational instance, the additional ML model,
determining, by the additional ML model, additional respective numerical similarities between the further textual record and the textual records in the additional corpus of textual records, and
based on the additional respective numerical similarities, transmitting, to the additional end-user computational instance, representations of one or more of the textual records in the additional corpus of textual records.

11. An article of manufacture including a non-transitory computer-readable medium, having stored thereon program instructions that, upon execution by a remote network management platform, cause the remote network management platform to perform operations comprising:
operating a first set of computational resources of the remote network management platform to provide an end-user computational instance that is dedicated to a managed network;
operating a second set of computational resources of the remote network management platform to provide a training computational instance that is configured to perform operations including: (i) receiving, from the end-user computational instance, a corpus of textual records, (ii) training, based on

Assignees

• Servicenow Inc

Inventors

• Jayaraman Baskar
• Thakur Aniruddha Madhusudhan
• Govindarajan Kannan
• Wong Andrew Kai Chiu
• Palapudi Sriram

Classifications

• G06N3/0499

  Feedforward networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• H04L67/34 Primary

  involving the movement of software or configuration parameters  (network booting or remote initial program loading [RIPL] G06F9/4416) · CPC title

• H04L41/16 Primary

  using machine learning or artificial intelligence · CPC title

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title