Abstract:
Although they take a lot of effort, modern code review procedures are crucial for preserving software quality. By utilizing Large Language Models (LLMs) and intelligent automation, AI-assisted code review solutions claim to increase reviewer productivity and find more flaws. In this study, we assess the impact of AI-assisted code reviews on code quality (defect identification, code enhancements) and developer productivity (e.g., review completion time, effort). We introduce a system that combines cutting-edge AI models (GPT-4, CodeBERT, CodeT5) with conventional static analysis tools.
కామెంట్లు లేవు:
కామెంట్ను పోస్ట్ చేయండి