Belief Function

Belief function theory, a framework for representing and manipulating uncertainty, aims to model incomplete or imprecise information, going beyond traditional probability by allowing for belief masses assigned to sets of possibilities rather than single events. Current research focuses on developing efficient algorithms for belief function computations, particularly in high-dimensional spaces, and integrating belief functions with other uncertainty models like probability and possibility theories, often within neural network architectures such as evidential neural networks or random-set convolutional neural networks. This approach finds applications in diverse fields, including machine learning (improving robustness and uncertainty quantification), multi-view clustering (handling imprecise data), and medical image segmentation (managing uncertainty in diagnoses).