Jun Tan

Title: Innovations of Jun Tan in Document Query Systems

Introduction

Jun Tan is an accomplished inventor based in Pittsburgh, PA. He has made significant contributions to the field of document retrieval systems. His innovative work focuses on enhancing the efficiency and effectiveness of querying rich documents using advanced language models.

Latest Patents

Jun Tan holds a patent for a groundbreaking invention titled "Hybrid retrieval augmented generation for rich document queries using a large language model." This patent outlines techniques for implementing hybrid retrieval augmented generation for rich document queries using a large language model (LLM). The method involves a computing system that receives documents added to databases and processes queries to generate tokens. It employs a probabilistic ranking method to determine the first ranking of documents and utilizes an embedded representation for a vector database query to establish a second ranking. The system then combines these rankings to select top documents and generates a knowledge graph based on them. Ultimately, the system outputs the query, the top documents, and a portion of the knowledge graph to the LLM, which produces a response.

Career Highlights

Jun Tan is currently employed at Zoom Communications, Inc., where he continues to innovate in the field of document retrieval. His work has garnered attention for its potential to revolutionize how users interact with large datasets and retrieve relevant information efficiently.

Collaborations

Jun Tan collaborates with talented individuals such as Shujian Liu and Ying Lu, who contribute to the development of advanced technologies in their respective fields.

Conclusion

Jun Tan's contributions to the field of document query systems exemplify the power of innovation in enhancing information retrieval. His patent and ongoing work at Zoom Communications, Inc. highlight his commitment to advancing technology for better user experiences.