The field of software engineering is witnessing a transformative shift with the integration of Large Language Models (LLMs) across various aspects of the development lifecycle. Recent advancements are notably enhancing code generation, refactoring, and testing processes, with LLMs demonstrating significant improvements in accuracy and efficiency. Personality-guided code generation and in-context learning are emerging as innovative strategies to tailor LLMs' outputs to specific coding tasks, thereby improving the quality and relevance of generated code. Additionally, the use of LLMs for automated software improvement and misconfiguration detection in serverless computing is gaining traction, offering more comprehensive and dynamic solutions compared to traditional methods. The potential of LLMs to generate executable oracles and automate the update of deprecated API usages further underscores their versatility and impact on modern software development practices. However, challenges such as the need for precise and complete code context to guide LLMs and the risk of unsafe refactorings highlight areas for future research and development. Overall, the integration of LLMs is not only reshaping how software is developed but also redefining the roles and responsibilities of software engineers in this evolving landscape.
LLMs Revolutionizing Software Engineering Practices
Sources
Personality-Guided Code Generation Using Large Language Models
Lingma SWE-GPT: An Open Development-Process-Centric Language Model for Automated Software Improvement
LLM-Based Misconfiguration Detection for AWS Serverless Computing
DemoCraft: Using In-Context Learning to Improve Code Generation in Large Language Models
LLMs: A Game-Changer for Software Engineers?
PairSmell: A Novel Perspective Inspecting Software Modular Structure
A Deep Dive Into Large Language Model Code Generation Mistakes: What and Why?
Generating executable oracles to check conformance of client code to requirements of JDK Javadocs using LLMs
Do Advanced Language Models Eliminate the Need for Prompt Engineering in Software Engineering?
Evaluating the Ability of Large Language Models to Generate Verifiable Specifications in VeriFast
An Empirical Study on the Code Refactoring Capability of Large Language Models
Parameter-Efficient Fine-Tuning of Large Language Models for Unit Test Generation: An Empirical Study
Utilizing Precise and Complete Code Context to Guide LLM in Automatic False Positive Mitigation
Interaction2Code: How Far Are We From Automatic Interactive Webpage Generation?
Automatic Generation of Question Hints for Mathematics Problems using Large Language Models in Educational Technology
Automated Update of Android Deprecated API Usages with Large Language Models
An Empirical Study on the Potential of LLMs in Automated Software Refactoring