Date: Wednesday, May 22
Start Time: 1:30 pm
End Time: 2:00 pm
Developers of machine-learning based computer vision applications often face difficulties obtaining sufficient data for training and evaluating models. In this talk, we explore the use of synthetic data techniques to overcome these challenges. We explain the “simulation-to-reality” gap and the challenges it poses for realistic synthetic data generation. Using an automated Mars exploration vehicle application, we share how we generate synthetic data using the Unity simulation environment, and demonstrate how synthetic data can be effective—while also highlighting its limitations. We also touch on the potential for few-shot learning techniques to reduce the need for training data.
