Teaching a Car to Drive Itself by Imitation and Imagination (Google I/O'19)

Опубликовано 8 мая 2019, 0:50

Training a neural net to drive by pure observation requires ensuring that good behavior is being learned from the right signals and that test results in simulation can be transferred to the real world. This talk will walk you through the evolution of a deep network to deal with these challenges by incorporating techniques like data perturbation, input and loss dropout, and domain-specific losses. You’ll learn how input ablation can help debug what the network has really learned.

Watch more #io19 here:

Google I/O 2019 All Sessions Playlist → goo.gle/io19allsessions
Learn more on the I/O Website → google.com/io

Subscribe to the Google Developers Channel → goo.gle/developers
Get started at → developers.google.com

Speaker(s): Mayank Bansal

T80E26 event: Google I/O 2019; re_ty: Publish; fullname: Mayank Bansal;