Abstract

Whilst mathematics has long been held as the language of science, computer code has rapidly become another such language; programming is now a critical tool in the scientist’s toolbox, not least for data analysis but also for the construction of complex models. The inherent and accidental complexity of such models can however stymie their use as a vehicle for both effectively communicating ideas and efficiently computing predictions. In this talk, I will discuss some general practical principles for ensuring that computer models can have more impact and value, both for yourself and the scientific community. I’ll use the established concept of FAI Rness (first applied to data): that computer models should be Findable, Accessible, Interoperable, and Reusable. To this I add my own perspective, that models should also be VRAI : Verifiable, Reproducible, Auditable, and Interpretable. Along the way I will talk about our work at ICCS applying such principles with earth and climate scientists and discuss our recent research on programming tools and systems to make it easier to develop FAIR and VRAI models.

———————

CAS UnionRoad is inviting you to a scheduled Zoom meeting.

Topic: CAS seminar: Dr Dominic Orchard Time: Nov 19, 2024 11:00 AM London

Join Zoom Meeting https://us02web.zoom.us/j/86167918145?pwd=iUbWnUiGzsG1qXb44kPI0Kkl64dsic.1

Meeting ID: 861 6791 8145 Passcode: 644297