Preface to the Special Issue on Software Quality Assurance in the Era of Large Language Models

In recent years, with the rapid advancement of deep learning technologies, large language models (LLMs) have achieved unprecedented breakthroughs in natural language processing.Representative models such as ChatGPT and DeepSeek are evolving rapidly and have demonstrated remarkable potential and broad application prospects in a range of software engineering tasks, including code generation, software testing, and program analysis and optimization.As general-purpose artificial intelligence models endowed with deep semantic understanding and generative capabilities, LLMs are redefining conventional approaches to software quality assurance, injecting new momentum into the automation and intelligence of software engineering.Additionally, software quality assurance plays a critical role in ensuring that software systems meet design specifications and user expectations in terms of functionality, performance, reliability, and security.However, conventional methods fail to keep pace with the increasing scale and complexity of modern software systems.This challenge is especially pronounced in highly complex application scenarios such as open-source software, enterpriselevel systems, and chip design programs, where the degree of automation and intelligence provided by conventional approaches often falls short of practical requirements.Driven by evolving technological demands and resonating with industrial developments, the research on the LLM-assisted software quality assurance is gaining increasing attention from both academia and industry.LLMs not only possess powerful capabilities in code comprehension and generation but also have achieved remarkable results in critical tasks such as software testing, code optimization, and code translation.Their large-scale pre-training paradigm and high scalability make them a key technological foundation for developing intelligent software quality assurance tools.Nevertheless, significant challenges remain in effectively integrating LLMs with existing software engineering practices to build intelligent systems that are applicable to real-world scenarios.