Quan Kong

Title: Quan Kong: Inventor of a Machine Learning-Based Search Ranking Patent.

Introduction

Quan Kong is an inventor based in Coquitlam, Canada. His patent work addresses how online search systems can rank new items when little or no historical interaction data is available.

Latest Patents

Kong is associated with 1 patent: “Machine learning-based ranking of cold-start items in search systems using predicted interaction rates.” The invention uses embeddings to identify established items similar to new items, then draws on those items’ historical performance to predict a baseline interaction rate. Search systems can use that prediction to rank or select new items for presentation to users. As interaction data accumulates for a new item, the system can blend the observed data with the predicted baseline until enough information is available.

Career Highlights

Kong’s listed patent is associated with Maplebear Inc. It focuses on applying machine learning and historical performance data to the challenge of ranking cold-start items.

Collaborations

Kong’s listed coworkers include Xiao Xiao and Lee Cohn.

Conclusion

Quan Kong’s patent describes a data-driven approach to ranking new items, using similarities to established items to help guide search results while direct interaction data is still limited.