Back to results
US 2026/0288734 A1Published application
System and method of an intelligent research engine
AssigneePiedmont Innovations, LLC
InventorsPaul Alexander Rudd, Marie Hazel George
Life of this patent
Estimating the term…
Abstract
Disclosed herein is a data processing engine that leverages a curated training dataset in combination with supplemental external data sources. This integration enables the system to operate asynchronously, generating predictive outputs tailored to user-specific parameters, even in the absence of active user engagement.
Claims20
A computer-implemented method comprising:
- populating a database based on a user objective;
- receiving a user prompt;
- querying the database based on the user prompt;
- retrieving structural and relational latent data from the database based on the user prompt, wherein the structural and relational latent data is relevant to the user prompt;
- providing relevant structural and relational latent data to a machine learning model, wherein the machine learning model outputs one or more latent user recommendations related to the user prompt;
- converting, using a large language model, the one or more latent user recommendations to one or more human-readable user recommendations;
- and
- displaying the one or more human-readable user recommendations.
5 more elements
A system comprising:
- a processor, and a memory storing instructions thereon, wherein execution of the instructions causes the processor to:
- populate a database based on a user objective;
- receive a user prompt;
- query the database based on the user prompt;
- retrieve structural and relational latent data from the database based on the user prompt, wherein the structural and relational latent data is relevant to the user prompt;
- provide relevant structural and relational latent data to a machine learning model, wherein the machine learning model outputs one or more latent user recommendations related to the user prompt;
- convert, using a large language model, the one or more latent recommendations to one or more human-readable user recommendations;
- and
- display the one or more human-readable user recommendations.
6 more elements
Dependencies are inferred from the claim text.