Capturing the flux in Scientific Knowledge

Scientific knowledge is in constant flux and our own understanding continually evolves to accommodate and explain new observations, instances and facts. The continuous change or flux is an important and inescapable feature of scientific knowledge; however, it's still not captured and utilized in contemporary science practices. The knowledge structures, such as taxonomies and databases that contemporary computational infrastructure use to represent scientific knowledge are static—their schemas are fixed so they cannot represent the fluid, dynamically changing nature of domain conceptualization. Example fields are biology—with an almost constant reorganization of the tree of life and geosciences, where the categories used to describe soils, rocks, landcover etc., often change in response to new scientific understanding or societal concerns. This research stems from the fluid process of science and endeavors to find ways to move from static and lifeless knowledge structures to dynamic, living and more meaningful knowledge structures to support our changing understanding.

This talk will present a model to represent dynamic categories enriched with social and contextual factors by connecting them to the process of their formation and revision. The natural sciences use categories to understand the world; a model that connects science processes and knowledge structures at different stages in the life cycle of a category would help sciences to understand and track the evolution of the world and our own understanding about the world. The proposed categorical model provides more bases to compare a category at different stages of its evolution, which would help to address the problem of information interoperability that is normally caused by semantic heterogeneity in databases. Similarly, it may also help to interoperate or harmonize science products, such as maps, that are based on categorical information. Such dynamic representation of knowledge would bring computational models closer to the real world models.

eResearch NZ 2013 session type: 


Submitted by Tim McNamara on