Adverse drug reactions (ADRs) significantly impact clinical medication safety. The timely identification and prediction of ADRs rely on the efficient analysis of real-world data, such as electronic health records, social media, and spontaneous reporting databases. In recent years, the rapid advancement of artificial intelligence, particularly large language models, in natural language processing, causal reasoning, and complex data mining has provided new technological means for real-time ADRs monitoring and individualized prediction. This paper summarizes the latest research achievements in AI-driven ADRs monitoring. Focusing on diverse data sources, including structured databases and electronic health records, it elaborates on the advantages andchallenges of AI in ADRs event extraction, relationship identification, causal analysis, and risk prediction. The aim is to provide a theoretical reference for constructing more intelligent and efficient ADRs monitoring systems.

Medical text representation is crucial for medical natural language processing (NLP) applications. Bidirectional encoder representations from transformers (BERT) has achieved the state-of-the-art performance in general domain text representation. However, limited by the design of the pretraining task and the frequency of knowledge occurrence, it lacks understanding of medical knowledge. To overcome these problems, we proposed a selective knowledge extraction and fusion framework to enhance medical text representation. In the knowledge extraction phase, we first designed a semantic importance evaluation metric to extract internal knowledge. We then used large language models (LLMs) to extract external knowledge from systematized nomenclature of medicine clinical term (SNOMED CT). In the knowledge fusion phase, we utilized an attention mechanism and Siamese network to integrate internal knowledge and external knowledge. Extracting knowledge through large language models (LLMs) and integrating it into five different types of BERT models, we achieved significant improvements in the task of pulmonary disease text classification.