Reproduce “Simulink Mutation Testing using CodeBERT” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jingfan Zhang, Delaram Ghobari, Mehrdad Sabetzadeh, Shiva Nejati
An approach named BERTiMuS adapts CodeBERT to generate mutants for Simulink models, demonstrating the application of Large Language Models to graphical languages. This method produces contextually informed faults, exhibiting superior performance in detecting requirements-relevant issues compared to a state-of-the-art Simulink mutation tool.
Key insights
- Large Language Models, specifically CodeBERT, can be effectively adapted for mutation testing of graphical modeling languages like Simulink by transforming their visual structure into a text-based format.
- The fine-tuned CodeBERT generates contextually-aware mutant instances that introduce realistic logical errors and can align with existing mutation patterns, going beyond simple syntactic changes.
- LLM-based mutant generation offers a complementary strategy to traditional rule-based tools, as each approach identifies distinct sets of test cases and reveals different types of faults.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.