Exploring full and parameter-efficient fine-tuning techniques for transformer-based design pattern detection

Ilyes Rezgui

doi:10.1016/j.infsof.2026.108176

Exploring full and parameter-efficient fine-tuning techniques for transformer-based design pattern detection

2026-05-15

0

PRE

AI

DOI：10.1016/j.infsof.2026.108176

Original

Original request for help

Save

Abstract

En 中文

Design patterns offer reusable solutions to common software development problems, improving code quality and maintainability. Despite their significance, design pattern instances in source code are often undocumented, making difficult for developers to identify and exploit them effectively. Automating pattern detection can support program comprehension, refactoring, and maintenance; however, existing approaches face several challenges.

Keywords:

design patterns

transformer models

fine-tuning techniques

software maintenance

pattern detection

Exploring full and parameter-efficient fine-tuning techniques for transformer-based design pattern detection

Abstract

Journal

Organization