|Wong, Ying Kit
|Transformer approach on source code vulnerability detection
|Computer software -- Quality control
Debugging in computer science
Hong Kong Polytechnic University -- Dissertations
|Faculty of Engineering
|1 volume (unpaged) : color illustrations
|Software quality is a major concern in the software business. Security attacks continue to escalate, leading to application failures, financial losses, and erosion of confidence. Therefore, it is critical to allow developers to discover security flaws in source code before deployment. Several static analysis techniques have been created to uncover security issues in source code. However, traditional approaches have not had a leveraged machine learning mechanism to reduce operational effort and improve detection performance.
Vulnerability identification is a crucial task in security. With the development of technology, identifying vulnerable source code has become a hot topic in the industry in order to handle large code bases. Applying natural language processing (NLP) to source code and building models to complete the analysis and digestion of source code has become one of the valuable vulnerability identification studies. Detecting errors before they cause damage is the desire of The solution. Previous efforts have usually been unsuccessful and inefficient.
This paper contributes an enhanced approach to detecting errors using the Transformer method in deep learning. The method is validated against a vulnerability benchmark database. We employ NLP to process source code as a sequence of text in a specific domain and embed the text by paying attention to a highly relevant portion of the code to solve the error detection problem with a pre-trained and then fine-tuned model.
|All rights reserved
Files in This Item:
|For All Users (off-campus access for PolyU Staff & Students only)
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: