Siemens Taps Omniverse Replicator for Synthetic Data Generation

The company aims to accelerate its AI model development times from taking “months” to “days”.

Siemens aims to accelerate its AI model development times by using NVIDIA Omniverse Replicator running on Amazon G5 instances for synthetic data generation. This should speed the workflow up from “months” to “days”, according to the company.

The Siemens Xcelerator and NVIDIA Omniverse platforms are joining to enable "full-design-fidelity, live digital twins that connect software-defined AI systems from edge to cloud."

Synthetic data, the information generated artificially as an alternative to real-world data, boosts data sets for robotics, safety monitoring, welding and wiring inspections, and checking kits of parts.

Using NVIDIA Replicator and Siemens SynthAI technology, we can procedurally generate sets of photorealistic images using the digital models of our products and production resources and an integrated training pipeline to train ready-to-use models. This speeds up our set-up time for AI inspection models by a factor of five,” said Maximilian Metzner, global lead for autonomous manufacturing systems for electronics at GWE.

Omniverse helps Siemens generate scenarios and more realistic images easily, aided with RTX technology-enabled physics-based rendering and materials. Siemens' SynthAI, designed to build data sets for training AI models, taps into Replicator and gets access to its randomization features "to vary the sizes and locations of defects, change lighting, color, texture and more to develop a robust dataset." Then, the data can be run through a defect detection model for training, which enables engineers to quickly test and iterate on models, with only a small set of data available at first.

Read more in NVIDIA's blog post and don't forget to join our Reddit page and our Telegram channel, follow us on Instagram and Twitter, where we share breakdowns, the latest news, awesome artworks, and more. 

Join discussion

Comments 0

    You might also like

    We need your consent

    We use cookies on this website to make your browsing experience better. By using the site you agree to our use of cookies.Learn more