Hypertexts combine rich knowledge structures with flexible ways of exploring them. Structured knowledge representations have been developed across domains ranging from the digital humanities to scientific research. Knowledge graphs have emerged as a particularly important framework for knowledge representation. However, most current knowledge graphs remain limited to establishing relatively simple facts. This work explores how knowledge graphs can contribute to the generation of interactive hypertexts—for example, how an enhanced Wikidata could more effectively support Wikipedia articles.

Knowledge Structures and Semantic Modeling One set of issues concerns the expressive power of knowledge graphs themselves. How can they support the description of events and complex scenarios? How can they accommodate the evolution of knowledge and alternative interpretations? And how can they enable semantic modeling—that is, the description of entire scenarios and even dynamic systems?

User Interaction A second set of issues surrounds user interaction with knowledge graphs and the hypertexts built upon them. Relevant capabilities include knowledge-graph visualization, argumentation and issue-based information systems, explanations, guided tours, and adaptive personalization. Potentially, the knowledge-graph application environment can incorporate discourse principles, session models, and user models.

Application Environments and Evaluation We will also consider novel forms of content, including multi-modal knowledge graphs, and examine applications across diverse domains such as digital humanities, eScience, healthcare, business, and government. Finally, we will consider methods for evaluating both the underlying graph structures and the quality of user interaction they support.