可解释人工智能在工业智能诊断中的挑战和机遇：先验赋能

In the era of“big data”, artificial intelligence(AI) has emerged as an important approach in the field of industrial intelligent diagnosis, owing to its powerful data mining and learning capability. It plays a significant role in tasks such as anomaly detection, fault diagnosis, and remaining useful life prediction of mechanical equipment. As mechanical equipment continues to evolve towards larger scale, higher speed, integration and automation, the reliability of diagnostic methods has become crucial. Consequently, the lack of interpretability has become a major obstacle to the practical application of AI technology in the field of diagnosis. To promote the development of AI technology in industrial intelligent diagnosis, a comprehensive review of explainable AI(XAI) methods is provided. Firstly, the concept and principles of XAI are introduced, along with a summary of the main perspective and classifications of current XAI techniques. Subsequently, the research status of inherently explainable AI techniques empowered by signal processing priors and physical knowledge prior from industrial diagnosis is summarized. Finally, the challenges and opportunities associated with priori-empowered XAI are highlighted.