I'm LongbridgeAI, I can summarize articles.Chip giant NVIDIA announced a collaboration with ride-hailing platform Uber to advance the development of autonomous driving technology.
On October 23, NVIDIA Drive posted on social media that it is exploring how to leverage large-scale real driving data to train foundational models to accelerate advancements in autonomous driving technology.
(NVIDIA Drive announces collaboration with Uber)
The post stated that NVIDIA will use the diverse driving scenario data collected by Uber—including airport pickups, complex intersections, and varying weather conditions—to perform post-training on NVIDIA's Cosmos World foundational model.
This collaboration relies on NVIDIA's DGX Cloud infrastructure and aims to achieve three major technical goals:
• Achieve higher precision in simulation, significantly reducing anomalies.
• Shorten the iteration cycle after model training, supporting scalable machine learning operations (MLOps).
• Ensure more stable and reliable model performance in rare or extreme scenarios.
Following the announcement, Uber's stock price rose to an intraday high on Thursday afternoon, with an increase of up to 3.5%, before narrowing its gains.
(Uber's stock price accelerated upward on Thursday)