Researchers have developed a new framework for automatically assigning and ordering International Classification of Diseases (ICD) codes in clinical notes. The LTR-ICD framework treats the task as a classification and ranking problem, considering the order of ICD codes. In experiments, the authors show that their approach outperforms previous methods, achieving higher accuracy in identifying high-priority codes and surpassing state-of-the-art models on classification metrics. --- Why it matters: This matters to researchers in AI because accurate assignment and ordering of ICD codes is essential for medical diagnosis and reimbursement. The proposed framework's ability to consider the order of codes could lead to more effective clinical decision support systems and improved patient care. Source: https://arxiv.org/abs/2510.13922