Recent developments in Artificial Intelligence (AI) have allowed machine learning models to solve certain complex problems in natural language processing and other areas at large scales by taking advantage of advancements in hardware and new neural network frameworks such as Recurrent Neural Networks and Transformers. We are now interested in synthesizing these advancements with other methods such as formal grammars, logic-based semantic representations, and knowledge graphs to improve the representation of knowledge, the evaluation of AI systems, and classification of documents and images. The use of semantically-derived taxonomies, ontologies, and relations automatically extracted from texts can improve the development and standardization of domain-based AI systems and human-computer interactions such as curation and information retrieval.