Method and system for generating text data
AssigneeThinkAnalytics Ltd.
InventorsChristopher McGuire, Peter Docherty
Life of this patent
Estimating the term…
Abstract
A computer-implemented method for generating text data comprising a descriptor for a group of content of items, the method comprising: obtaining content metadata associated with a group of content items; generating a descriptor for the group of content items based on at least the content metadata associated with the group; at least one of displaying and storing the generated descriptor.
Claims19
A computer-implemented method for generating text data comprising a descriptor for a group of content of items, the method comprising:
- obtaining, from at least a first data source, content metadata associated with a group of content items, wherein the content metadata is represented by at least one feature vector or other mathematical representation;
- processing, by a processing resource, the at least one feature vector or other mathematical representation to generate a prompt for a descriptive text generator model;
- transmitting, via a model interface, a request over a network to the model server, wherein the request comprises the generated prompt;
- generating, by the descriptive text generator model at the model server, a descriptor for the group of content items in response to receiving the request, based on the prompt;
- receiving, by the processing resource, the generated descriptor from the model server;
- at least one of displaying, by a display, and storing, by a storage resource, the generated descriptor.
(canceled)
A system comprising a processing circuitry configured to generate text data comprising a descriptor for a group of content of items, wherein the processing circuitry is configured to:
- obtain content metadata, from at least one data source, associated with a group of content items, wherein the content metadata is represented by at least one feature vector or other mathematical representation;
- process, by the processing resource, the at least one feature vector or other mathematical representation to generate a prompt for a descriptive text generator model;
- transmitting, via a model interface, a request over a network to the model server, wherein the request comprises the generated prompt;
- generate a descriptor for the group of content items based on at least the content metadata associated with the group;
- and
- receiving, by the processing resource, the generated descriptor from the model server;
- wherein the system comprises at least one of:
- a display configured to display the generated descriptor and a storage resource for storing the generated descriptor.
(canceled)
A non-transitory computer-readable medium that comprises computer-readable instructions that, when executed by a processor, cause the processor to:
- obtain content metadata, form at least one data source, associated with a group of content items, wherein the content metadata is represented by at least one feature vector or other mathematical representation;
- process, by the processing resource, the at least one feature vector or other mathematical representation to generate a prompt for a descriptive text generator model;
- transmit, via a model interface, a request over a network to the model server, wherein the request comprises the generated prompt;
- generate a descriptor for the group of content items based on at least the content metadata associated with the group;
- receiving, by the processing resource, the generated descriptor from the model server;
- and
- at least one of transmit the generated descriptor to a display for display or store the generated descriptor.
Dependencies are inferred from the claim text.