Department of Electrical Engineering and Computer Science
Christopher S. Bond Life Sciences Center
University of Missouri, Columbia, MO, 65201, USA
To address the problems of scalar neurons, a novel deep-learning architecture, known as Capsule Network (CapsNet) was introduced in 2017. The main building block of CapsNet is the capsule, which is a group of neuron vectors whose lengths represent the entity probabilities. Capsule provides a unique and powerful deep-learning building block to better model the diverse relationships inside internal representations of a neural network. We have applied CapsNet in several applications and achieved improved performance over previous deep-learning methods. In this talk, I will discuss the concept, implementation and applications of CapsNet in details.