On-device agentic validation via personalized model ensemble
AssigneeU.S. Bank National Association
InventorRahul Sharma
Life of this patent
Estimating the term…
Abstract
An agentic recommendation is generated in response to a query from an agentic model executing in a user device. Further in response to the query, a set of personalized recommendations is generated from a set of personalized models forming an ensemble of models executing in the user device. In an output corresponding to the query, the agentic recommendation is replaced with at least one personalized recommendation from the set of personalized recommendations.
Claims20
A method, comprising:
- executing locally on a user device, a plurality of models comprising an agentic model and an ensemble of models, the ensemble of models comprising a set of personalized models, wherein the agentic model is delivered to the user device pre-trained with unchanging weights, and wherein each personalized model in the set of personalized model is trained locally on the user device on a different function of an application present on the user device;
- generating, from the agentic model, and responsive to a query, an agentic recommendation;
- generating, from the ensemble of models, and responsive to the query, a set of personalized recommendations;
- and
- replacing, in an output corresponding to the query, the agentic recommendation with at least one personalized recommendation from the set of personalized recommendations.
A computer program product comprising a set of one or more non-transitory computer readable storage media, and program instructions collectively stored on the one or more non-transitory computer readable storage media, the program instructions executable by a set of one or more processors to cause the set of one or more processors to perform operations comprising:
- executing locally on a user device, a plurality of models comprising an agentic model and an ensemble of models, the ensemble of models comprising a set of personalized models, wherein the agentic model is delivered to the user device pre-trained with unchanging weights, and wherein each personalized model in the set of personalized model is trained locally on the user device on a different function of an application present on the user device;
- generating, from the agentic model, and responsive to a query, an agentic recommendation;
- generating, from the ensemble of models, and responsive to the query, a set of personalized recommendations;
- and
- replacing, in an output corresponding to the query, the agentic recommendation with at least one personalized recommendation from the set of personalized recommendations.
A system comprising a set of one or more processors and one or more non-transitory computer readable storage media, and program instructions collectively stored on the one or more non-transitory computer readable storage media, the program instructions executable by the set of one or more processors to cause the processor to perform operations comprising:
- executing locally on a user device, a plurality of models comprising an agentic model and an ensemble of models, the ensemble of models comprising a set of personalized models, wherein the agentic model is delivered to the user device pre-trained with unchanging weights, and wherein each personalized model in the set of personalized model is trained locally on the user device on a different function of an application present on the user device;
- generating, from the agentic model, and responsive to a query, an agentic recommendation;
- generating, from the ensemble of models, and responsive to the query, a set of personalized recommendations;
- and
- replacing, in an output corresponding to the query, the agentic recommendation with at least one personalized recommendation from the set of personalized recommendations.
Dependencies are inferred from the claim text.