Invited Speak at the Workshop on Autonomous Driving at CVPR 2020, see http://cvpr2020.wad.eyesight
Summary:
Machine learning is important to establishing a self-driving stack that can scale to a diverse established of environments without the need of demanding exhaustive guide labeling or professional tuning. In this converse, I will describe some of the latest modeling get the job done at Waymo that aims to capture far better the inherent framework in the autonomous driving domain. I will give an update on the Waymo Open up Dataset and the just lately completed Worries and emphasize some of our the latest perform on self-supervision of notion models, as very well as details-driven techniques for sensor simulation.
Bio:
Drago joined Waymo in 2018 to lead the Exploration group, which focuses on creating the condition of the art in autonomous driving using machine learning. Drago invested 8 many years at Google to start with working on 3D vision and pose estimation for StreetView, and later leading a investigation crew which formulated computer vision units for annotating Google Pics. The staff also invented common solutions these kinds of as the Inception neural network architecture, and the SSD detector, which served earn the Imagenet 2014 Classification and Detection problems. Ahead of Waymo, Drago led the 3D Notion staff at Zoox.
