Talking Autonomy: Training and Testing KineticFlow

In this episode Ghost Autonomy co-founder and CTO Volkmar Uhlig explains how KineticFlow - Ghost's core visual perception AI algorithm - is trained and tested. KineticFlow's universal, physics-based approach to perception offers many inherent benefits in training, testing, and validation. In this video, Volkmar covers the training and validation pipeline: how Ghost collects ground truth data (video) in its test cars; how data is auto-labeled in the data center without any manual human labeling; and how training data sets are curated to ensure proper coverage of every hyperspace dimension while avoiding overfitting. Volkmar also details how this process is automated to ensure efficient training and re-training as Ghost adds capabilities and support for new sensors, vehicles, and automakers. Thank you for watching Talking Autonomy, a series of short tech talks that explain the key elements of our work at Ghost. Subscribe and stay tuned for new episodes as we continue exploring the core technologies behind the Ghost Autonomy Engine and sharing insights from the founders, engineers, designers, mathematicians, and even the policy-makers who are responsible for bringing Ghost’s self-driving technology to the roads. Connect with Ghost and learn more: - Ghost Autonomy Engine on the web: https://www.ghostautonomy.com/platform - KineticFlow vision neural network: https://www.ghostautonomy.com/kineticflow - Follow Ghost on Twitter: https://twitter.com/ghostautonomy - Follow Ghost on LinkedIn: https://www.linkedin.com/company/GhostAutonomy #selfdriving #AI #autonomous #GhostAutonomy #computervision