Database systems with user-configurable automated metadata assignment

Patent metadata

Field	Value
Publication number	US-12499319-B2
Application number	US-202218064724-A
Country	US
Kind code	B2
Filing date	Dec 12, 2022
Priority date	Dec 12, 2022
Publication date	Dec 16, 2025
Grant date	Dec 16, 2025

Abstract

Database systems and methods are provided for assigning structural metadata to records and creating automations using the structural metadata. One method of assigning structural metadata to a group of conversation records involves receiving a user input modification pertaining to a group of semantically similar conversations, automatically reassigning a conversation to a different group of semantically similar conversations based on its representative utterance in a manner that is influenced by the user input modification, and automatically updating, at a database system, a record associated with the conversation to include metadata identifying the different group of semantically similar conversations.

First claim

What is claimed is:

1 . A method comprising:
receiving user input including a modification pertaining to at least one of a plurality of preexisting groups of semantically similar conversations, a respective group of the plurality of preexisting groups comprising a subset of conversations previously assigned to the respective group based on a respective representative utterance associated with the respective conversation of the subset of conversations; automatically reassigning a first conversation of the subset of conversations from the respective group to a different group of semantically similar conversations based on the representative utterance associated with the first conversation in a manner that is influenced by the modification using a user-constrained assignment algorithm comprising one or more constraints corresponding to the modification; automatically updating, at a database system, a record associated with the first conversation to include metadata identifying the different group of semantically similar conversations in lieu of the respective group previously assigned to the first conversation; determining a value for a performance metric associated with the different group of the plurality of preexisting groups after automatically reassigning the first conversation, wherein the value for the performance metric is influenced by one or more performance metrics associated with the first conversation and an updated value for the performance metric associated with the respective group is influenced by reassignment of the first conversation; and providing one or more graphical indicia influenced by the value of the performance metric associated with the different group and the updated value for the performance metric associated with the respective group.

2 . The method of claim 1 , wherein:
the modification comprises a new group to be added to the plurality of preexisting groups of semantically similar conversations; automatically reassigning the first conversation comprises automatically reassigning the first conversation from a preexisting group of the plurality of preexisting groups of semantically similar conversations to the new group based on the representative utterance associated with the first conversation; and automatically updating the record comprises automatically updating the record to include an updated group value for a first field of the metadata identifying the new group in lieu of a previous group value for the first field of the metadata identifying the preexisting group.

3 . The method of claim 2 , wherein:
the user input includes a semantic representation assigned to the new group; and automatically reassigning the first conversation comprises automatically reassigning the first conversation based on a relationship between the semantic representation assigned to the new group and the representative utterance associated with the first conversation.

4 . The method of claim 3 , further comprising:
generating a first numerical representation of the representative utterance associated with the first conversation; and generating a second numerical representation of the semantic representation assigned to the new group, wherein automatically reassigning the first conversation comprises automatically reassigning the conversation to the new group when a difference between the first numerical representation and the second numerical representation is less than a second difference between the first numerical representation and a third numerical representation of a second semantic representation assigned to the preexisting group.

5 . The method of claim 1 , wherein:
the modification comprises a second conversation to be assigned to the different group of the plurality of preexisting groups; and automatically reassigning the first conversation comprises automatically reassigning the first conversation from a preexisting group of the plurality of preexisting groups of semantically similar conversations to the different group based on a relationship between the representative utterance associated with the first conversation and the respective representative utterance associated with the second conversation.

6 . The method of claim 1 , wherein:
the modification comprises removal of a first group of the plurality of preexisting groups of semantically similar conversations; and automatically reassigning the first conversation comprises automatically reassigning the first conversation from the first group of the plurality of preexisting groups of semantically similar conversations to the different group based on the representative utterance associated with the first conversation.

7 . The method of claim 1 , wherein:
the modification comprises merging a first group of the plurality of preexisting groups of semantically similar conversations with a second group of the plurality of preexisting groups of semantically similar conversations; and automatically reassigning the first conversation comprises automatically reassigning the first conversation from the first group of the plurality of preexisting groups of semantically similar conversations to the second group based on the representative utterance associated with the first conversation.

8 . The method of claim 1 , further comprising configuring one or more parameters of a Gaussian mixture model (GMM) using the modification to obtain a semi-supervised GMM constrained by the user input, wherein automatically reassigning the first conversation comprises:
inputting a numerical representation of the representative utterance associated with the first conversation to the semi-supervised GMM to obtain an output from the semi-supervised GMM influenced by the numerical representation of the representative utterance associated with the first conversation; and reassigning the first conversation from a first group of the plurality of preexisting groups of semantically similar conversations to a second group of the plurality of preexisting groups of semantically similar conversations when the output from the semi-supervised GMM is indicative of the second group.

9 . The method of claim 8 , wherein each group of the plurality of preexisting groups is distinct relative to other groups of the plurality of preexisting groups.

10 . At least one non-transitory machine-readable storage medium that provides instructions that, when executed by at least one processor, are configurable to cause the at least one processor to perform operations comprising:
receiving user input including a modification pertaining to at least one of a plurality of preexisting groups of semantically similar conversations, a respective group of the plurality of preexisting groups comprising a subset of conversations previously assigned to the respective group based on a respective representative utterance associated with the respective conversation of the subset of conversations; automatically reassigning a first conversation of the subset of conversations from the respective group to a different group of semantically similar conversations based on the representative utterance associated with the first conversation in a manner that is influenced by the modification using a user-constrained assignment algorithm comprising one or more constraints corresponding to the modification; automatically updating a record associated with the first conversation to include metadata identifying the different group of semantically similar conversations in lieu of the respective group previously assigned to the first conversation; determining a value for a performance metric associated with the different group of the plurality of preexisting groups after automatically reassignin

Assignees

• Salesforce Inc

Inventors

• Alexander Zachary
• Mao Yixin

Classifications

• G06F16/164 Primary

  File meta data generation · CPC title

• G06F40/35 Primary

  Discourse or dialogue representation · CPC title