20240028740. SOURCE CODE VULNERABILITY DETECTION AND REPAIR THROUGH MACHINE LEARNING simplified abstract (MICROSOFT TECHNOLOGY LICENSING, LLC.)

From WikiPatents
Jump to navigation Jump to search

SOURCE CODE VULNERABILITY DETECTION AND REPAIR THROUGH MACHINE LEARNING

Organization Name

MICROSOFT TECHNOLOGY LICENSING, LLC.

Inventor(s)

AARON YUE-CHIU Chan of PROVO UT (US)

COLIN BRUCE Clement of SEATTLE WA (US)

YEVHEN Mohylevskyy of REDMOND WA (US)

NEELAKANTAN Sundaresan of BELLEVUE WA (US)

ROSHANAK Zilouchian Moghaddam of KIRKLAND WA (US)

SOURCE CODE VULNERABILITY DETECTION AND REPAIR THROUGH MACHINE LEARNING - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240028740 titled 'SOURCE CODE VULNERABILITY DETECTION AND REPAIR THROUGH MACHINE LEARNING

Simplified Explanation

The abstract of the patent application describes a system that uses a neural classifier model to detect cybersecurity vulnerabilities in source code. This classifier model is trained on source code that may contain security bugs. If the classifier model identifies a source code snippet as likely containing a cybersecurity vulnerability, a proposed repair for the vulnerability is predicted using a neural decoder transformer model. This decoder model is trained on non-vulnerable source code and is used to generate source code that fixes the identified vulnerability.

  • The system uses a neural classifier model to detect cybersecurity vulnerabilities in source code.
  • The classifier model is trained on source code that may contain security bugs.
  • If a source code snippet is classified as likely containing a vulnerability, a repair for the vulnerability is predicted.
  • The repair is generated using a neural decoder transformer model.
  • The decoder model is trained on non-vulnerable source code.
  • The generated source code serves as a fix for the identified cybersecurity vulnerability.

Potential applications of this technology:

  • Automated detection and repair of cybersecurity vulnerabilities in source code.
  • Enhancing the security of software systems by identifying and fixing vulnerabilities.
  • Assisting developers in writing secure code by providing automated vulnerability detection and repair.

Problems solved by this technology:

  • Manual detection of cybersecurity vulnerabilities in source code can be time-consuming and error-prone.
  • Identifying and fixing vulnerabilities in large codebases can be challenging and resource-intensive.
  • This technology automates the process of vulnerability detection and provides potential fixes, saving time and effort.

Benefits of this technology:

  • Improved security of software systems by proactively identifying and fixing vulnerabilities.
  • Increased efficiency in vulnerability detection and repair processes.
  • Assistance to developers in writing secure code and reducing the risk of cybersecurity breaches.


Original Abstract Submitted

a neural classifier model is used to detect cybersecurity vulnerabilities in the source code predicted by a deep learning code generation model having been trained on source code possibly containing security bugs. upon the classifier model classifying a given source code snippet as likely containing a cybersecurity vulnerability, a proposed repair for the cybersecurity vulnerability is predicted from a neural decoder transformer model having been trained on non-vulnerable source code. the neural decoder transformer model is used to predict source code that repairs the cybersecurity vulnerability given the source code classified with a cybersecurity vulnerability.

Retrieved from "http://wikipatents.org/index.php?title=20240028740._SOURCE_CODE_VULNERABILITY_DETECTION_AND_REPAIR_THROUGH_MACHINE_LEARNING_simplified_abstract_(MICROSOFT_TECHNOLOGY_LICENSING,_LLC.)&oldid=30040"