Using synthetic datasets to train artificial intelligence (AI) is a fast, cost-effective way to deploy robust embedded algorithms for new hardware technologies. In this webinar by onscale, we demonstrate how an embedded AI algorithm for 3D smartphone touchscreens was trained and validated using the results of 8,000 simulations run in parallel. The AI can predict the location of a touch on the phone’s screen using the outputs of simple MEMS strain sensors on the underside of the glass. In addition to location, the AI provides information on the forces being applied to the glass, providing users with a further dimension in which to interact with their device. Synthetic datasets generated from cloud simulations can be created in hours using the OnScale platform, compared to weeks or months to create similar datasets from physical experimentation. The following topics will be covered: – Overview of the problem statement – Review of the methodology used to run the simulations and train the AI – Demo of the AI in operation: use of machine learning (ML) techniques to update Digital Twins – Q&A session
Presentor: Andrew Tweedie. Email: andrew.tweedie@onscale.com
Register here.