Fault Localization

Fault localization aims to automatically identify the root cause of errors in complex systems, ranging from software programs and cloud infrastructure to deep neural networks and hydroelectric units. Current research focuses on developing efficient algorithms, including those leveraging machine learning models like large language models (LLMs), graph convolutional networks (GCNs), and ensemble methods, to pinpoint faulty components or code segments, often addressing challenges like imbalanced data and limited labeled samples. These advancements are crucial for improving software reliability, accelerating debugging processes, and enhancing the trustworthiness of AI systems in various applications. The development of new datasets specifically designed to avoid data leakage issues in LLM-based approaches is also a significant area of focus.