Chen Qu

Title: Chen Qu and His Patent on Efficient Dual-Encoder Retrieval.

Introduction

Chen Qu is an inventor based in Sunnyvale, California, and works at Google Inc. His work includes a patent on improving retrieval efficiency in dual-encoder systems. The supplied information lists 1 patent for him.

Latest Patents

“Dual Encoder Retrieval Efficiency with Parameter Sharing in Projection Layer” describes an asymmetric dual-encoder architecture for processing two inputs, such as a question and an answer. Separate token-embedding and encoder sections process each input, while a shared projection layer receives both encodings and generates projections. These projections support an embedding space that produces question and answer embeddings, which can be used to identify candidate answers to an input query.

Career Highlights

Qu’s listed patent focuses on parameter sharing in the projection layer, an approach intended to support efficient retrieval while retaining separate processing paths for the two inputs. The patent information provided identifies Google Inc. as his company.

Collaborations

The supplied information names Zhe Dong and Jianmo Ni as Qu’s coworkers. Their inclusion here reflects the provided coworker information and does not establish their roles in the patent.

Conclusion

Chen Qu’s listed invention applies a shared projection layer to an asymmetric dual-encoder architecture, supporting question-and-answer embeddings for candidate-answer retrieval.