Jiayuan Zhou

Title: Jiayuan Zhou: Advancing Software Vulnerability Detection

Introduction

Jiayuan Zhou is an inventor based in Kista, Sweden. Her patent work addresses the detection and reporting of software vulnerability fixes in code repositories. She is associated with Huawei Technologies Co., Ltd.

Latest Patents

Zhou is named on one patent, “Systems and Methods for Detection of Software Vulnerability Fix.” The invention describes a method for analyzing code changes in a repository to predict whether a commit fixes a vulnerability. It preprocesses a commit into file-level token sequences, uses a transformer model to create an embedding for each file’s code changes, and combines those embeddings into a commit-level representation. A classifier then generates a predicted vulnerability-fix score, which is included in a report identifying the commit. Zhou’s listed patent portfolio comprises 1 patent.

Career Highlights

The patent reflects work at the intersection of software security and machine learning. By applying transformer-based representations and classification to repository commits, the described system aims to help identify potentially important vulnerability fixes.

Collaborations

Zhou’s listed coworkers include Michael Pacheco and Xin Xia. Her association with Huawei Technologies Co., Ltd. provides the organizational context for her patent work.

Conclusion

Jiayuan Zhou’s patent focuses on using machine learning to detect and report code changes that may resolve software vulnerabilities.