Systems and methods for identifying relevant data to use for enriching data structures in large datasets using detected metadata correlations
AssigneeCapital One Services, LLC
InventorsMark Merling, Tania Cruz Morales, Joshua Edwards
Life of this patent
Estimating the term…
Abstract
Systems and methods for uses and/or improvements to using artificial intelligence applications with large datasets. In particular, systems and methods for enriching data structures (e.g., data structures comprising user profile data) with additional data structures (and values thereof) from accessible data from a plurality of other users based on identifying relevant data to use for enriching the data structures using detected metadata correlations.
Claims20
A system for identifying relevant data to use for enriching data structures in large datasets using detected metadata correlations, the system comprising:
- one or more processors;
- and
- a non-transitory, computer-readable medium having instructions recorded thereon that, when executed by the one or more processors, cause operations comprising:
- receiving user data available from a first source, wherein the user data comprises a first data structure with a plurality of metadata types, and wherein the user data includes a first metadata type of the plurality of metadata types and a first value corresponding to the first metadata type;
- receiving a user input requesting a recommendation based on the user data;
- determining a second metadata type for enriching the user data that is required to generate the recommendation, wherein the plurality of metadata types does not include the second metadata type;
- filtering accessible data for a plurality of users based on the first metadata type to generate a first subset of the accessible data that corresponds to the first metadata type;
- selecting, based on the first metadata type, a first model of a plurality of models for extracting data from the first subset, wherein each of the plurality of models comprises a respective extraction algorithm, and wherein each respective extraction algorithm determines one or more correlations between the first metadata type and the second metadata type;
- generating a first input to the first model based on the first subset and the first value;
- processing the first input in the first model to generate a first output based on the one or more correlations between the first metadata type and the second metadata type, wherein the first output comprises a second value corresponding to the second metadata type;
- and
- generating a second data structure for the user data by enriching the first data structure with the second value.
A method for identifying relevant data to use for enriching data structures in large datasets using detected metadata correlations, the method comprising:
- receiving user data, wherein the user data comprises a first data structure with a plurality of metadata types, and wherein the user data includes a first metadata type of the plurality of metadata types and a first value corresponding to the first metadata type;
- determining a second metadata type for enriching the user data, wherein the plurality of metadata types does not include the second metadata type;
- filtering accessible data for a plurality of users based on the first metadata type to generate a first subset of the accessible data that corresponds to the first metadata type;
- selecting, based on the first metadata type, a first model of a plurality of models for extracting data from the first subset, wherein each of the plurality of models comprises a respective extraction algorithm;
- generating a first input to the first model based on the first subset and the first value;
- processing the first input in the first model to generate a first output, wherein the first output comprises a second value corresponding to the second metadata type;
- and
- generating a second data structure for the user data by enriching the first data structure with the second value.
A non-transitory, computer readable medium storing instructions that, when executed by one or more processors, cause operations comprising:
- receiving user data, wherein the user data comprises a first data structure with a plurality of metadata types, and wherein the user data includes a first metadata type of the plurality of metadata types and a first value corresponding to the first metadata type;
- filtering accessible data for a plurality of users based on the first metadata type to generate a first subset of the accessible data that corresponds to the first metadata type;
- selecting, based on the first metadata type, a first model of a plurality of models for extracting data from the first subset, wherein each of the plurality of models comprises a respective extraction algorithm;
- generating a first input to the first model based on the first subset and the first value;
- processing the first input in the first model to generate a first output, wherein the first output comprises a second value corresponding to a second metadata type;
- and
- generating a second data structure for the user data by enriching the first data structure with the second value.
Dependencies are inferred from the claim text.