Cognitive Discovery is an overarching framework that relies on AI to achieve scientific knowledge extraction and representation, to intelligently design and guide simulations, in order to drastically accelerate the pace of scientific discovery and technical R&D. Figure 2 illustrates the framework. Cognitive Discovery targets to massively accelerate scientific workflows in highly technical disciplines and provide a new generation of tools. Such workflows typically on the following cycle: a) Massive and tedious literature review in order to understand the problem at hand and formulate a methodological solution. It is key to point out that literature refers to all aspects such as mathematical modeling and (numerical) solution methods, actual computer models and HPC deployment aspects to name a few. b) Enrichment of literature data with experimental evidence and formation of hypotheses in order to search a configuration space. c) Running simulations to test hypotheses and generate new knowledge in order to close any knowledge gaps and progress the search. We argue that all three phases suffer today major disruptions. Simply put: 1) the volume of new literature in all technical fields is exploding (e.g. roughly 450K new publications in materials science are published every year, tens of thousands of papers in numerical and HPC methods need to be reviewed). 2) IoT advances as well as advances in measuring all aspects of HPC systems (and the applications on these) create an explosion of data. 3) High fidelity models lead to massive configuration spaces the complexity of which clearly outpaces our capability to scale and efficiently run massively parallel heterogeneous HPC systems. We will showcase how AI can help dramatically improve this setting and lead to a massive acceleration for scientific discovery.
Back to Workshop II: HPC and Data Science for Scientific Discovery