Physical simulations are at the core of many critical industrial systems. However, today's physical simulators have some limitations such as computation time, dealing with missing or uncertain data, or even non-convergence for some feasible cases.
Recently, the use of ML techniques to learn complex physical simulations has been considered as a promising approach to address those issues. However, this comes often at the cost of some accuracy which may hinder the industrial use. To drive this new research topic towards a better real-world applicability, we propose to discuss recent advances in designing augmented physical simulation, especially methodologies and techniques related to their validation towards an industrial deployment.
The validation of augmented physical simulation with regard to both ML and industrial criteria, is a key tenet of their success in operation and requires leveraging a span of criteria including statistical significance, physical compliance, industrial readiness, among others.
Objectives of the workshop
This workshop will be the place to discuss R&D subjects related to the above-mentioned challenges, including but not limited to:
- Evaluation of deep learning techniques-based physical simulation
- End-to-End benchmarking tools that accommodate augmented physical simulations
- Deployment of AI-based simulations in industry
- Talk 1: state of the art on AI-Based Simulation
- Talk2: AI-Based simulation in power grid industry
- Talk3: Benchmarking framework for evaluating AI-based simulators