Jun Qian

Title: Jun Qian: Innovator in Machine Learning and Document Processing

Introduction

Jun Qian is a prominent inventor based in Pasadena, CA, known for his contributions to machine learning and document processing technologies. With a total of 8 patents, he has made significant strides in developing automated techniques for generating synthetic training data and improving machine learning models.

Latest Patents

One of Jun Qian's latest patents focuses on generating synthetic training data, specifically document images with key-value pairs. This innovative approach allows for the automated generation of a large volume of diverse training data, which can be utilized to train machine learning models for extracting key-value pairs from document images. By using a single input document image and associated annotation data, the synthetic data generation system creates numerous synthetic training datapoints, each containing a synthetic document image and relevant annotation data. This advancement enhances the performance of machine learning models in extracting key-value pairs from document images.

Another notable patent involves multi-stage machine learning model training for key-value extraction. This technique outlines a systematic approach to train a machine learning model using a set of training data that includes unlabeled documents from various categories. The initial training stage focuses on identifying relationships among tokens, such as words, numbers, and punctuation in documents. Subsequently, the model is re-trained using a specific category of documents, while excluding others. The second stage is a supervised machine learning phase, where the training data is labeled to identify key-value pairs in the documents. This structured training process optimizes the model's parameters based on the characteristics of the training dataset.

Career Highlights

Jun Qian is currently employed at Oracle International Corporation, where he continues to innovate and develop cutting-edge technologies in the field of machine learning and document processing. His work has significantly impacted the efficiency and accuracy of data extraction processes.

Collaborations

Jun has collaborated with notable colleagues, including Iman Zadeh and Tao Sheng, contributing to various projects that enhance the capabilities of machine learning applications.

Conclusion

Jun Qian's innovative work in machine learning and document processing has led to significant advancements in the field. His patents reflect a commitment to improving the efficiency of data extraction and training processes, making him a valuable contributor to the technology landscape.