Bellevue, WA, United States of America

Xinying Song

This inventor holds 16 USPTO granted patents and 6 published patent applications and 1 EPO patent, primarily in Machine Learning. Top assignees: Microsoft Technology Licensing, LLC, Google Inc.. Active years: 2015-2026.

IDiyas Innovation Intelligence. (2026). Inventor Profile: Xinying Song. Retrieved from https://idiyas.com/inventor/xinying-song

Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile

Calculating rank
USPTO Granted Patents = 16 

% Patents Active = 100.0


Average Co-Inventor Count = 5.3

ph-index = 5

Forward Citations = 51(Granted Patents)


Location History:

  • Harbin, CN (2015)
  • Redmond, WA (US) (2020 - 2022)
  • Bellevue, WA (US) (2016 - 2023)

Company Filing History:


Years Active: 2015-2026

Loading Chart...
Loading Chart...
Areas of Expertise:
Large Language Models
Tokenization
Resource Scheduling
Machine Learning
Task Routing
Workflow Engine
Intelligent Scheduling
Computational Models
Contextual Recommendations
Data Record Extraction
Relevance Group Suggestions
Salient Item Identification
16 patents (USPTO):Explore Patents

Title: The Innovative Mind of Xinying Song

Introduction

Xinying Song, based in Bellevue, WA, is a prominent inventor recognized for his contributions to the field of technology with a remarkable portfolio of 14 patents. His inventions focus on advancing computational methods, particularly in the realms of language processing and resource management.

Latest Patents

Among Xinying Song's latest patents is the development of inference methods for word or wordpiece tokenization. This innovation introduces systems and methods that leverage a left-to-right longest-match-first greedy process. The approach organizes vocabulary into a trie structure, allowing efficient parsing in a single pass and enhancing the tokenizer's ability to generate a list of tokens corresponding to the longest matching vocabulary entries. Another significant patent pertains to resource scheduling using machine learning. This method involves identifying task features and utilizing a gradient boost tree model to predict the likelihood of successful task completion, ultimately allowing for optimized scheduling of tasks to be accomplished by various resources.

Career Highlights

Xinying Song has worked with some of the industry’s leading companies, including Microsoft Technology Licensing, LLC and Google Inc. His diverse experience in these innovative environments has undoubtedly contributed to his expansive knowledge and skill set in developing cutting-edge technologies.

Collaborations

Throughout his career, Xinying has collaborated with several notable professionals, including Jianfeng Gao and Hui Su. These collaborations have fostered an environment of shared ideas and advancements, further enriching his contributions to technology.

Conclusion

Xinying Song’s work exemplifies the spirit of innovation. With his extensive patent portfolio and impactful collaborations, he continues to push the boundaries of technology, making significant strides in areas that directly influence how we interact with machines and manage resources. His contributions are an inspiring part of the ongoing technological evolution.

Profile summary based on public USPTO records.
Data Sources: USPTO Patent Grant XML, Patent Center, EPO & CIPO • Normalized by IDiyas Innovation Graph. Methodology & provenance architecturePlease report any incorrect information to [email protected]
Loading…