Jolene Huang

Title: Jolene Huang and Retrieval-Augmented Generation

Introduction

Jolene Huang is an inventor based in Olympia, Washington, and is associated with Intuit, Inc. Her patent work addresses retrieval-augmented generation, a technique that helps language models produce responses informed by relevant data.

Latest Patents

Huang is named on 1 patent, titled “Retrieval augmented generation.” The patent describes storing data embeddings in an embedding database and retrieving relevant embeddings when a related prompt is received. Those embeddings are used to create an augmented prompt, which is then provided to a machine-learning model. By incorporating information from the database, the approach can improve the breadth and depth of the model’s responses, even when the model was not trained on the underlying data.

Career Highlights

Huang’s listed invention focuses on connecting external data sources with language-model responses through embedding retrieval and prompt augmentation. This approach offers a way to make model outputs more relevant to specific information without requiring the model itself to have prior knowledge of that data.

Collaborations

Huang’s listed coworkers include Pankaj Rastogi and Spriha Awasthi. The patent information associates her work with Intuit, Inc.

Conclusion

Jolene Huang’s patent on retrieval-augmented generation describes a method for enriching language-model responses with relevant information retrieved from an embedding database.