WAABI Innovation Inc. (20250103779). LEARNING UNSUPERVISED WORLD MODELS FOR AUTONOMOUS DRIVING VIA DISCRETE DIFFUSION
LEARNING UNSUPERVISED WORLD MODELS FOR AUTONOMOUS DRIVING VIA DISCRETE DIFFUSION
Organization Name
Inventor(s)
Sergio Casas Romero of Toronto CA
LEARNING UNSUPERVISED WORLD MODELS FOR AUTONOMOUS DRIVING VIA DISCRETE DIFFUSION
This abstract first appeared for US patent application 20250103779 titled 'LEARNING UNSUPERVISED WORLD MODELS FOR AUTONOMOUS DRIVING VIA DISCRETE DIFFUSION
Original Abstract Submitted
a method learns unsupervised world models for autonomous driving via discrete diffusion. the method includes encoding an observation of an actor for a geographic region using an encoder to generate a prior frame of prior tokens. the method further includes processing the prior frame with a spatio-temporal transformer to generate a predicted frame of predicted tokens. the spatio-temporal transformer includes a spatial transformer and a temporal transformer. the method further includes processing the predicted frame to generate a predicted action for the actor. the method further includes decoding the predicted frame to generate a predicted observation of the geographic region.