Severity Classification

Severity classification aims to automatically assign levels of disease severity or event impact from various data sources, such as medical images, sensor readings, or textual reports. Current research focuses on developing and refining machine learning models, including convolutional neural networks, vision transformers, and Bayesian active learning approaches, to improve accuracy and efficiency, often addressing challenges like imbalanced datasets and the need for efficient annotation strategies. This work has significant implications for improving diagnostic accuracy, enabling personalized treatment, optimizing resource allocation (e.g., in healthcare or traffic management), and facilitating more effective risk assessment across diverse fields.

Papers