Intuit, Inc.

John Patrick Symborski

Siddharth Jain

Sivashanker Thiruchittampalam

Rachit Debashish Sengupta

Jianzhao Huang

A method includes constructing a schema segment from a schema segment definition. The schema segment includes a multitude of schema elements. A multitude of embeddings corresponding to the multitude of schema elements is generated by an embedding model. The multitude of schema elements and the multitude of embeddings is grouped into a multitude of categories. A semantic collision between a first embedding of a first schema element and a second embedding of a second schema element of a category of the multitude of categories may be detected. A set of semantic collisions may be obtained. A large language model (LLM) may resolve the set of semantic collisions to obtain a corresponding multitude of updated schema elements. The schema segment is updated with the corresponding multitude of updated schema elements.

Schema generation with collision detection and correction

A method includes obtaining, as a response to a user query, a multitude of search results and corresponding similarity scores. The method further includes obtaining a multitude of data sources corresponding to the multitude of search results. The method further includes processing a multi-metric analytic function for the search results, and corresponding data sources to obtain a composite score for the search results. The method further includes selecting a subset of search results from the multitude of search results based on a composite score threshold to obtain a selected subset of search results. The method further includes selecting a corresponding subset of data sources from the multitude of data sources to obtain a selected subset of data sources. The method further includes generating a prompt to a large language model (LLM), including at least the user query, the selected subset of search results, and the selected subset of data sources.

Growing community of inventors

Edmonton, Canada

John Patrick Symborski