Next-Generation Neural Networks: Capsule Networks With Routing-by-Agreement for Text Classification

These days, neural networks constantly prove their high capacity for nearly every application case and are considered as key technology learning systems. However, need to continuously evolve managing new arising challenges like increasing task complexity, explainability of decision making processes, expanded problem domains, providing resilient robust systems etc. One possible enhancement traditional constitutes the innovative Capsule Network (CapsNet) technology, which combines expressiveness distributed entity representations with an intelligent interpretable signal propagation, named routing-by-agreement. Since CapsNets represent a relatively young acquirement, further research is essential gaining profound knowledge about CapsNet theory best practices diverse areas. This paper wants contribute progress text classification. For this purpose, various questions get formulated experimentally answered aid six selected datasets. In addition, serves starting point researchers well practitioners deal in domain, by supplying survey its theory, classification basics combination both The analysis results empirically robustness routing-by-agreement wide spectrum net architectures, datasets tasks. Hence, can be viewed next-generation network offers potential method should topic future research.