Abstract. Scientific progress comes from adapting and extending models from prior work to address new problems. However, this ideal workflow is difficult primarily because of the informal or inconsistent representation of models. This problem is exacerbated in fields outside of computer science where scientific models are formulated in informal languages. We propose SemanticModels.jl, a category theory-based framework for defining meta-modeling tasks such as model augmentation and model selection along with semantic information extraction. We illustrate the major features of SemanticModels.jl including representating models as wiring diagrams, extending and composing models with algebraic operations, and generating executable code of the resulting models. These features are demonstrated by constructing a model of mosquito borne illness in humans along with predator-prey dynamics between mosquitos and birds to study the effects of predator species on the control of mosquito borne illnesses. Through the careful application of category theoretic ideas to scientific computing, general patterns in scientific modeling languages and frameworks can be axiomatized and these formalizations can be exploited to improve scientific computing research and development processes.
| PP20 Home 2020 | Program | Speaker Index |