Abstract
Software vulnerability detection is increasingly important as modern applications combine multiple programming
languages. This paper presents an early comparative evaluation of BERT, RoBERTa, and CodeBERT for binary vulnerability
detection across HTML, Python, JavaScript, and PHP using the CVEFixes dataset and language-wise three-fold stratified
cross-validation. The results show clear performance differences across languages, indicating that multilingual
vulnerability detection requires more language-aware and robust transformer-based modelling strategies.
languages. This paper presents an early comparative evaluation of BERT, RoBERTa, and CodeBERT for binary vulnerability
detection across HTML, Python, JavaScript, and PHP using the CVEFixes dataset and language-wise three-fold stratified
cross-validation. The results show clear performance differences across languages, indicating that multilingual
vulnerability detection requires more language-aware and robust transformer-based modelling strategies.
| Original language | English |
|---|---|
| Title of host publication | SPECS Conference 2026 |
| Pages | 1-4 |
| Number of pages | 4 |
| Publication status | Published - 30 May 2026 |
Fingerprint
Dive into the research topics of 'Early Comparative Evaluation of Transformer Models for Multilingual Software Vulnerability Detection'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver