Setting a Unified Standard for AI Model Reproducibility & Validation

Scientists create new standards for AI models

Aspiring bakers often have to adapt winning recipes for different kitchen configurations. For example, someone might make chocolate chip cookies using an eggbeater rather than a stand mixer.

It is important for talented chefs to be able to replicate a recipe under different conditions and with different setups. Computational scientists also face the same problem when they are trying to validate or work with AI models. These models are used in a variety of scientific fields, from brain research to climate analysis.

\”When we speak about data, we can have a good understanding of what we are dealing with,\” said Eliu Huerta, a scientist at the Argonne National Lab of the U.S. Department of Energy (DOE). With an AI model it is less clear. Are we talking about smartly structured data, computing, software, or a combination?