Wayve, a London-based AI startup, has developed a machine-learning model capable of driving multiple types of vehicles, including passenger cars and delivery vans. This achievement marks the first time a single AI driver has been able to adapt to different vehicle types. Initially, the model was trained to drive cars using thousands of hours of driving data. The team then trained the AI to drive vans in a simulation, requiring only 80 hours of additional data, which significantly improved the model's adaptability. The approach used by Wayve combines reinforcement learning, where the AI learns through trial and error, and imitation learning, where it mimics human driving behaviors. This method allows the AI to generalize its driving skills across different vehicles more efficiently than previously expected.
The
investment will also support Wayve in scaling its foundation models, advancing
research in embodied AI, and expanding its operations and partnerships
globally. This funding is seen as a significant boost for the European AI
sector, reflecting strong investor confidence in AI's potential to
revolutionize autonomous driving and other industries.
No comments:
Post a Comment