[Advances in the application of artificial intelligence in antimicrobial resistance research]

Antimicrobial resistance (AMR) has emerged as a major threat to global public health, while conventional research methods face severe bottlenecks in deciphering its complex mechanisms and accelerating new drug development. Artificial intelligence (AI), particularly deep learning, is revolutionizing AMR research by enabling the processing of high-dimensional multi-omics data, uncovering hidden patterns, and generating novel hypotheses. This review systematically elaborates on the biomedical big data ecosystem that drives the AI revolution, including multi-omics data, phenotypic and clinical data, and literature-based knowledge data. We then discuss in detail cutting-edge AI methods and their applications in multi-level resistance mechanism analysis (knowledge-enhanced retrieval, resistance gene identification, and phenotype prediction) and intelligent design of novel antimicrobial molecules. Furthermore, we analyze core challenges in data quality, algorithm interpretability, clinical translation, and ethical governance. Finally, we propose key future directions, such as building equitable data ecosystems, developing interpretable AI models, and deepening interdisciplinary collaborations. This review aims to provide researchers with a comprehensive perspective on the current landscape, existing challenges, and future paths for AI applications in the AMR field.