The Use of Synthetic Data in AI model Training

AI Spectrum

11-05-2022 • 21 mins

Creating an accurate AI model requires millions of images and data points to be fed into the computing system. This is a difficult task that can slow down the speed to market or lower the accuracy of the model that is created. Synthetic data helps in solving this problem by reducing the amount of real data that needs to be collected. That results in reduced time to market and increased model accuracy. In today’s episode, I’m talking to Zachi Mann. He leads a new initiative that is focused on advanced robotics simulation capabilities at Siemens. He’ll help us understand AI model training for factory robots. He’ll also share with us how Siemens solutions such as CAD and NX help in model development. In this episode, you’ll learn about the use of synthetic data in training AI-reliant factory robots. You’ll also learn about the challenges and the benefits that come with combining synthetic data with real data. Additionally, you’ll learn about Synth AI, a new synthetic data generating solution from Siemens. In this episode, you will learn: The meaning of synthetic data (03:03) The challenges that come with the use of synthetic data (04:19) The benefits of using synthetic data (08:05) Why the use of synthetic data has been on the rise (11:06) Other uses of synthetic data besides AI model training (17:25) Connect with Zachi Mann: LinkedIn Connect with Spencer Acain: LinkedIn