Wisconsin COSMOS Use Case Narrative:
I’m a geoscientist (true!). I have an old Fortran model cGENIE that does an ok job of modeling the coupled biogeochemical components of the surface earth (carbon focused). The model runs and makes predictions about the state of things in the world, but it needs to be parameterized and validated. This REQUIRES data and observations from the field. These data are very difficult to get.
I bring this model code to the new ASKE facility. The model code is parsed and key entities (equations, variables, comment descriptions of phenomena etc) are extracted (will ASKE rewrite it for me and/or make it better?). The UW COSMOS facility from ASKE is used to “intersect” this model (code, input, and output) with the scientific literature to identify a collection of scientific documents spanning all domains of science (could be MILLIONS) that are pertinent to the model and its parts. Documents with tables, figures, equations, text within them are linked to specific code components. Tables containing data are identified and those that are relevant to model predictions are located, extracted, parsed, and data in them aggregated for use in model parameterization and/or validation. New phenomena and data pertinent to model and its predictions are discovered. COSMOS HTC infrastructure and document acquisition pipelines (a facility that is part of a broader ASKE infrastructure) make it possible to get results quickly. Updated and new results are provided every day as the literature continues to be published and as UW COSMOS acquires some 10K new documents daily.
Replace “old Fortran model” above with just about anything: a manually-constructed database, a collection of scientific papers, a question stated in plain language… our goal is to be able to fully wield the entirety of published scientific knowledge in a way that will enhance, if not completely transform, our ability to use, parameterize, construct, and assess scientific models, augment and extend mission-critical scientific databases, and improve our ability to leverage past investments in data acquisition and publication to enable new science.
crapo
commented
5 years ago
Wisconsin COSMOS Use Case Narrative: I’m a geoscientist (true!). I have an old Fortran model cGENIE that does an ok job of modeling the coupled biogeochemical components of the surface earth (carbon focused). The model runs and makes predictions about the state of things in the world, but it needs to be parameterized and validated. This REQUIRES data and observations from the field. These data are very difficult to get.
I bring this model code to the new ASKE facility. The model code is parsed and key entities (equations, variables, comment descriptions of phenomena etc) are extracted (will ASKE rewrite it for me and/or make it better?). The UW COSMOS facility from ASKE is used to “intersect” this model (code, input, and output) with the scientific literature to identify a collection of scientific documents spanning all domains of science (could be MILLIONS) that are pertinent to the model and its parts. Documents with tables, figures, equations, text within them are linked to specific code components. Tables containing data are identified and those that are relevant to model predictions are located, extracted, parsed, and data in them aggregated for use in model parameterization and/or validation. New phenomena and data pertinent to model and its predictions are discovered. COSMOS HTC infrastructure and document acquisition pipelines (a facility that is part of a broader ASKE infrastructure) make it possible to get results quickly. Updated and new results are provided every day as the literature continues to be published and as UW COSMOS acquires some 10K new documents daily.
Replace “old Fortran model” above with just about anything: a manually-constructed database, a collection of scientific papers, a question stated in plain language… our goal is to be able to fully wield the entirety of published scientific knowledge in a way that will enhance, if not completely transform, our ability to use, parameterize, construct, and assess scientific models, augment and extend mission-critical scientific databases, and improve our ability to leverage past investments in data acquisition and publication to enable new science.